MyArxiv
Computation and Language 90
☆ A Grounded Typology of Word Classes
We propose a grounded approach to meaning in language typology. We treat data from perceptual modalities, such as images, as a language-agnostic representation of meaning. Hence, we can quantify the function--form relationship between images and captions across languages. Inspired by information theory, we define "groundedness", an empirical measure of contextual semantic contentfulness (formulated as a difference in surprisal) which can be computed with multilingual multimodal language models. As a proof of concept, we apply this measure to the typology of word classes. Our measure captures the contentfulness asymmetry between functional (grammatical) and lexical (content) classes across languages, but contradicts the view that functional classes do not convey content. Moreover, we find universal trends in the hierarchy of groundedness (e.g., nouns > adjectives > verbs), and show that our measure partly correlates with psycholinguistic concreteness norms in English. We release a dataset of groundedness scores for 30 languages. Our results suggest that the grounded typology approach can provide quantitative evidence about semantic function in language.
comment: 19 pages, 5 figures
☆ AdvPrefix: An Objective for Nuanced LLM Jailbreaks
Many jailbreak attacks on large language models (LLMs) rely on a common objective: making the model respond with the prefix "Sure, here is (harmful request)". While straightforward, this objective has two limitations: limited control over model behaviors, often resulting in incomplete or unrealistic responses, and a rigid format that hinders optimization. To address these limitations, we introduce AdvPrefix, a new prefix-forcing objective that enables more nuanced control over model behavior while being easy to optimize. Our objective leverages model-dependent prefixes, automatically selected based on two criteria: high prefilling attack success rates and low negative log-likelihood. It can further simplify optimization by using multiple prefixes for a single user request. AdvPrefix can integrate seamlessly into existing jailbreak attacks to improve their performance for free. For example, simply replacing GCG attack's target prefixes with ours on Llama-3 improves nuanced attack success rates from 14% to 80%, suggesting that current alignment struggles to generalize to unseen prefixes. Our work demonstrates the importance of jailbreak objectives in achieving nuanced jailbreaks.
☆ SCBench: A KV Cache-Centric Analysis of Long-Context Methods
Long-context LLMs have enabled numerous downstream applications but also introduced significant challenges related to computational and memory efficiency. To address these challenges, optimizations for long-context inference have been developed, centered around the KV cache. However, existing benchmarks often evaluate in single-request, neglecting the full lifecycle of the KV cache in real-world use. This oversight is particularly critical, as KV cache reuse has become widely adopted in LLMs inference frameworks, such as vLLM and SGLang, as well as by LLM providers, including OpenAI, Microsoft, Google, and Anthropic. To address this gap, we introduce SCBench(SharedContextBench), a comprehensive benchmark for evaluating long-context methods from a KV cachecentric perspective: 1) KV cache generation, 2) KV cache compression, 3) KV cache retrieval, 4) KV cache loading. Specifically, SCBench uses test examples with shared context, ranging 12 tasks with two shared context modes, covering four categories of long-context capabilities: string retrieval, semantic retrieval, global information, and multi-task. With it, we provide an extensive KV cache-centric analysis of eight categories long-context solutions, including Gated Linear RNNs, Mamba-Attention hybrids, and efficient methods such as sparse attention, KV cache dropping, quantization, retrieval, loading, and prompt compression. The evaluation is conducted on 8 long-context LLMs. Our findings show that sub-O(n) memory methods suffer in multi-turn scenarios, while sparse encoding with O(n) memory and sub-O(n^2) pre-filling computation perform robustly. Dynamic sparsity yields more expressive KV caches than static patterns, and layer-level sparsity in hybrid architectures reduces memory usage with strong performance. Additionally, we identify attention distribution shift issues in long-generation scenarios. https://aka.ms/SCBench.
☆ Interlocking-free Selective Rationalization Through Genetic-based Learning
A popular end-to-end architecture for selective rationalization is the select-then-predict pipeline, comprising a generator to extract highlights fed to a predictor. Such a cooperative system suffers from suboptimal equilibrium minima due to the dominance of one of the two modules, a phenomenon known as interlocking. While several contributions aimed at addressing interlocking, they only mitigate its effect, often by introducing feature-based heuristics, sampling, and ad-hoc regularizations. We present GenSPP, the first interlocking-free architecture for selective rationalization that does not require any learning overhead, as the above-mentioned. GenSPP avoids interlocking by performing disjoint training of the generator and predictor via genetic global search. Experiments on a synthetic and a real-world benchmark show that our model outperforms several state-of-the-art competitors.
☆ DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding
We present DeepSeek-VL2, an advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL, through two key major upgrades. For the vision component, we incorporate a dynamic tiling vision encoding strategy designed for processing high-resolution images with different aspect ratios. For the language component, we leverage DeepSeekMoE models with the Multi-head Latent Attention mechanism, which compresses Key-Value cache into latent vectors, to enable efficient inference and high throughput. Trained on an improved vision-language dataset, DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition, document/table/chart understanding, and visual grounding. Our model series is composed of three variants: DeepSeek-VL2-Tiny, DeepSeek-VL2-Small and DeepSeek-VL2, with 1.0B, 2.8B and 4.5B activated parameters respectively. DeepSeek-VL2 achieves competitive or state-of-the-art performance with similar or fewer activated parameters compared to existing open-source dense and MoE-based models. Codes and pre-trained models are publicly accessible at https://github.com/deepseek-ai/DeepSeek-VL2.
☆ Still "Talking About Large Language Models": Some Clarifications
My paper "Talking About Large Language Models" has more than once been interpreted as advocating a reductionist stance towards large language models. But the paper was not intended that way, and I do not endorse such positions. This short note situates the paper in the context of a larger philosophical project that is concerned with the (mis)use of words rather than metaphysics, in the spirit of Wittgenstein's later writing.
☆ One world, one opinion? The superstar effect in LLM responses
As large language models (LLMs) are shaping the way information is shared and accessed online, their opinions have the potential to influence a wide audience. This study examines who the LLMs view as the most prominent figures across various fields, using prompts in ten different languages to explore the influence of linguistic diversity. Our findings reveal low diversity in responses, with a small number of figures dominating recognition across languages (also known as the "superstar effect"). These results highlight the risk of narrowing global knowledge representation when LLMs retrieve subjective information.
☆ Benchmarking Linguistic Diversity of Large Language Models
The development and evaluation of Large Language Models (LLMs) has primarily focused on their task-solving capabilities, with recent models even surpassing human performance in some areas. However, this focus often neglects whether machine-generated language matches the human level of diversity, in terms of vocabulary choice, syntactic construction, and expression of meaning, raising questions about whether the fundamentals of language generation have been fully addressed. This paper emphasizes the importance of examining the preservation of human linguistic richness by language models, given the concerning surge in online content produced or aided by LLMs. We propose a comprehensive framework for evaluating LLMs from various linguistic diversity perspectives including lexical, syntactic, and semantic dimensions. Using this framework, we benchmark several state-of-the-art LLMs across all diversity dimensions, and conduct an in-depth case study for syntactic diversity. Finally, we analyze how different development and deployment choices impact the linguistic diversity of LLM outputs.
☆ Reasoner Outperforms: Generative Stance Detection with Rationalization for Social Media
Stance detection is crucial for fostering a human-centric Web by analyzing user-generated content to identify biases and harmful narratives that undermine trust. With the development of Large Language Models (LLMs), existing approaches treat stance detection as a classification problem, providing robust methodologies for modeling complex group interactions and advancing capabilities in natural language tasks. However, these methods often lack interpretability, limiting their ability to offer transparent and understandable justifications for predictions. This study adopts a generative approach, where stance predictions include explicit, interpretable rationales, and integrates them into smaller language models through single-task and multitask learning. We find that incorporating reasoning into stance detection enables the smaller model (FlanT5) to outperform GPT-3.5's zero-shot performance, achieving an improvement of up to 9.57%. Moreover, our results show that reasoning capabilities enhance multitask learning performance but may reduce effectiveness in single-task settings. Crucially, we demonstrate that faithful rationales improve rationale distillation into SLMs, advancing efforts to build interpretable, trustworthy systems for addressing discrimination, fostering trust, and promoting equitable engagement on social media.
☆ Targeted Angular Reversal of Weights (TARS) for Knowledge Removal in Large Language Models
The sheer scale of data required to train modern large language models (LLMs) poses significant risks, as models are likely to gain knowledge of sensitive topics such as bio-security, as well the ability to replicate copyrighted works. Methods designed to remove such knowledge must do so from all prompt directions, in a multi-lingual capacity and without degrading general model performance. To this end, we introduce the targeted angular reversal (TARS) method of knowledge removal from LLMs. The TARS method firstly leverages the LLM in combination with a detailed prompt to aggregate information about a selected concept in the internal representation space of the LLM. It then refines this approximate concept vector to trigger the concept token with high probability, by perturbing the approximate concept vector with noise and transforming it into token scores with the language model head. The feedforward weight vectors in the LLM which operate directly on the internal representation space, and have the highest cosine similarity with this targeting vector, are then replaced by a reversed targeting vector, thus limiting the ability of the concept to propagate through the model. The modularity of the TARS method allows for a sequential removal of concepts from Llama 3.1 8B, such as the famous literary detective Sherlock Holmes, and the planet Saturn. It is demonstrated that the probability of triggering target concepts can be reduced to 0.00 with as few as 1 TARS edit, whilst simultaneously removing the knowledge bi-directionally. Moreover, knowledge is shown to be removed across all languages despite only being targeted in English. Importantly, TARS has minimal impact on the general model capabilities, as after removing 5 diverse concepts in a modular fashion, there is minimal KL divergence in the next token probabilities of the LLM on large corpora of Wikipedia text (median of 0.002).
comment: 14 pages, 5 figures, 1 table
☆ Efficient Continual Pre-training of LLMs for Low-resource Languages
Open-source Large Language models (OsLLMs) propel the democratization of natural language research by giving the flexibility to augment or update model parameters for performance improvement. Nevertheless, like proprietary LLMs, Os-LLMs offer poorer performance on low-resource languages (LRLs) than high-resource languages (HRLs), owing to smaller amounts of training data and underrepresented vocabulary. On the other hand, continual pre-training (CPT) with large amounts of language-specific data is a costly proposition in terms of data acquisition and computational resources. Our goal is to drastically reduce CPT cost. To that end, we first develop a new algorithm to select a subset of texts from a larger corpus. We show the effectiveness of our technique using very little CPT data. In search of further improvement, we design a new algorithm to select tokens to include in the LLM vocabulary. We experiment with the recent Llama-3 model and nine Indian languages with diverse scripts and extent of resource availability. For evaluation, we use IndicGenBench, a generation task benchmark dataset for Indic languages. We experiment with various CPT corpora and augmented vocabulary size and offer insights across language families.
☆ How good is my story? Towards quantitative metrics for evaluating LLM-generated XAI narratives
A rapidly developing application of LLMs in XAI is to convert quantitative explanations such as SHAP into user-friendly narratives to explain the decisions made by smaller prediction models. Evaluating the narratives without relying on human preference studies or surveys is becoming increasingly important in this field. In this work we propose a framework and explore several automated metrics to evaluate LLM-generated narratives for explanations of tabular classification tasks. We apply our approach to compare several state-of-the-art LLMs across different datasets and prompt types. As a demonstration of their utility, these metrics allow us to identify new challenges related to LLM hallucinations for XAI narratives.
☆ Retrieval-Augmented Semantic Parsing: Using Large Language Models to Improve Generalization
Open-domain semantic parsing remains a challenging task, as models often rely on heuristics and struggle to handle unseen concepts. In this paper, we investigate the potential of large language models (LLMs) for this task and introduce Retrieval-Augmented Semantic Parsing (RASP), a simple yet effective approach that integrates external lexical knowledge into the parsing process. Our experiments not only show that LLMs outperform previous encoder-decoder baselines for semantic parsing, but that RASP further enhances their ability to predict unseen concepts, nearly doubling the performance of previous models on out-of-distribution concepts. These findings highlight the promise of leveraging large language models and retrieval mechanisms for robust and open-domain semantic parsing.
comment: Submitted to ARR
☆ VLR-Bench: Multilingual Benchmark Dataset for Vision-Language Retrieval Augmented Generation COLING 2025
We propose the VLR-Bench, a visual question answering (VQA) benchmark for evaluating vision language models (VLMs) based on retrieval augmented generation (RAG). Unlike existing evaluation datasets for external knowledge-based VQA, the proposed VLR-Bench includes five input passages. This allows testing of the ability to determine which passage is useful for answering a given query, a capability lacking in previous research. In this context, we constructed a dataset of 32,000 automatically generated instruction-following examples, which we denote as VLR-IF. This dataset is specifically designed to enhance the RAG capabilities of VLMs by enabling them to learn how to generate appropriate answers based on input passages. We evaluated the validity of the proposed benchmark and training data and verified its performance using the state-of-the-art Llama3-based VLM, the Llava-Llama-3 model. The proposed VLR-Bench and VLR-IF datasets are publicly available online.
comment: The 31st International Conference on Computational Linguistics (COLING 2025), 19 pages
☆ TACOMORE: Leveraging the Potential of LLMs in Corpus-based Discourse Analysis with Prompt Engineering
The capacity of LLMs to carry out automated qualitative analysis has been questioned by corpus linguists, and it has been argued that corpus-based discourse analysis incorporating LLMs is hindered by issues of unsatisfying performance, hallucination, and irreproducibility. Our proposed method, TACOMORE, aims to address these concerns by serving as an effective prompting framework in this domain. The framework consists of four principles, i.e., Task, Context, Model and Reproducibility, and specifies five fundamental elements of a good prompt, i.e., Role Description, Task Definition, Task Procedures, Contextual Information and Output Format. We conduct experiments on three LLMs, i.e., GPT-4o, Gemini-1.5-Pro and Gemini-1.5.Flash, and find that TACOMORE helps improve LLM performance in three representative discourse analysis tasks, i.e., the analysis of keywords, collocates and concordances, based on an open corpus of COVID-19 research articles. Our findings show the efficacy of the proposed prompting framework TACOMORE in corpus-based discourse analysis in terms of Accuracy, Ethicality, Reasoning, and Reproducibility, and provide novel insights into the application and evaluation of LLMs in automated qualitative studies.
☆ ROUTE: Robust Multitask Tuning and Collaboration for Text-to-SQL
Despite the significant advancements in Text-to-SQL (Text2SQL) facilitated by large language models (LLMs), the latest state-of-the-art techniques are still trapped in the in-context learning of closed-source LLMs (e.g., GPT-4), which limits their applicability in open scenarios. To address this challenge, we propose a novel RObust mUltitask Tuning and collaboration mEthod (ROUTE) to improve the comprehensive capabilities of open-source LLMs for Text2SQL, thereby providing a more practical solution. Our approach begins with multi-task supervised fine-tuning (SFT) using various synthetic training data related to SQL generation. Unlike existing SFT-based Text2SQL methods, we introduced several additional SFT tasks, including schema linking, noise correction, and continuation writing. Engaging in a variety of SQL generation tasks enhances the model's understanding of SQL syntax and improves its ability to generate high-quality SQL queries. Additionally, inspired by the collaborative modes of LLM agents, we introduce a Multitask Collaboration Prompting (MCP) strategy. This strategy leverages collaboration across several SQL-related tasks to reduce hallucinations during SQL generation, thereby maximizing the potential of enhancing Text2SQL performance through explicit multitask capabilities. Extensive experiments and in-depth analyses have been performed on eight open-source LLMs and five widely-used benchmarks. The results demonstrate that our proposal outperforms the latest Text2SQL methods and yields leading performance.
☆ Can LLMs Convert Graphs to Text-Attributed Graphs?
Graphs are ubiquitous data structures found in numerous real-world applications, such as drug discovery, recommender systems, and social network analysis. Graph neural networks (GNNs) have become a popular tool to learn node embeddings through message passing on these structures. However, a significant challenge arises when applying GNNs to multiple graphs with different feature spaces, as existing GNN architectures are not designed for cross-graph feature alignment. To address this, recent approaches introduce text-attributed graphs, where each node is associated with a textual description, enabling the use of a shared textual encoder to project nodes from different graphs into a unified feature space. While promising, this method relies heavily on the availability of text-attributed data, which can be difficult to obtain in practice. To bridge this gap, we propose a novel method named Topology-Aware Node description Synthesis (TANS), which leverages large language models (LLMs) to automatically convert existing graphs into text-attributed graphs. The key idea is to integrate topological information with each node's properties, enhancing the LLMs' ability to explain how graph topology influences node semantics. We evaluate our TANS on text-rich, text-limited, and text-free graphs, demonstrating that it enables a single GNN to operate across diverse graphs. Notably, on text-free graphs, our method significantly outperforms existing approaches that manually design node features, showcasing the potential of LLMs for preprocessing graph-structured data, even in the absence of textual information. The code and data are available at https://github.com/Zehong-Wang/TANS.
☆ ASLoRA: Adaptive Sharing Low-Rank Adaptation Across Layers
As large language models (LLMs) grow in size, traditional full fine-tuning becomes increasingly impractical due to its high computational and storage costs. Although popular parameter-efficient fine-tuning methods, such as LoRA, have significantly reduced the number of tunable parameters, there is still room for further optimization. In this work, we propose ASLoRA, a cross-layer parameter-sharing strategy combining global sharing with partial adaptive sharing. Specifically, we share the low-rank matrix A across all layers and adaptively merge matrix B during training. This sharing mechanism not only mitigates overfitting effectively but also captures inter-layer dependencies, significantly enhancing the model's representational capability. We conduct extensive experiments on various NLP tasks, showing that ASLoRA outperforms LoRA while using less than 25% of the parameters, highlighting its flexibility and superior parameter efficiency. Furthermore, in-depth analyses of the adaptive sharing strategy confirm its significant advantages in enhancing both model flexibility and task adaptability.
☆ Familiarity: Better Evaluation of Zero-Shot Named Entity Recognition by Quantifying Label Shifts in Synthetic Training Data
Zero-shot named entity recognition (NER) is the task of detecting named entities of specific types (such as 'Person' or 'Medicine') without any training examples. Current research increasingly relies on large synthetic datasets, automatically generated to cover tens of thousands of distinct entity types, to train zero-shot NER models. However, in this paper, we find that these synthetic datasets often contain entity types that are semantically highly similar to (or even the same as) those in standard evaluation benchmarks. Because of this overlap, we argue that reported F1 scores for zero-shot NER overestimate the true capabilities of these approaches. Further, we argue that current evaluation setups provide an incomplete picture of zero-shot abilities since they do not quantify the label shift (i.e., the similarity of labels) between training and evaluation datasets. To address these issues, we propose Familiarity, a novel metric that captures both the semantic similarity between entity types in training and evaluation, as well as their frequency in the training data, to provide an estimate of label shift. It allows researchers to contextualize reported zero-shot NER scores when using custom synthetic training datasets. Further, it enables researchers to generate evaluation setups of various transfer difficulties for fine-grained analysis of zero-shot NER.
comment: 8 pages, 4 figures, 5 tables
☆ Label-template based Few-Shot Text Classification with Contrastive Learning
As an algorithmic framework for learning to learn, meta-learning provides a promising solution for few-shot text classification. However, most existing research fail to give enough attention to class labels. Traditional basic framework building meta-learner based on prototype networks heavily relies on inter-class variance, and it is easily influenced by noise. To address these limitations, we proposes a simple and effective few-shot text classification framework. In particular, the corresponding label templates are embed into input sentences to fully utilize the potential value of class labels, guiding the pre-trained model to generate more discriminative text representations through the semantic information conveyed by labels. With the continuous influence of label semantics, supervised contrastive learning is utilized to model the interaction information between support samples and query samples. Furthermore, the averaging mechanism is replaced with an attention mechanism to highlight vital semantic information. To verify the proposed scheme, four typical datasets are employed to assess the performance of different methods. Experimental results demonstrate that our method achieves substantial performance enhancements and outperforms existing state-of-the-art models on few-shot text classification tasks.
☆ MALAMUTE: A Multilingual, Highly-granular, Template-free, Education-based Probing Dataset
Language models (LMs) have excelled in various broad domains. However, to ensure their safe and effective integration into real-world educational settings, they must demonstrate proficiency in specific, granular areas of knowledge. Existing cloze-style benchmarks, commonly used to evaluate LMs' knowledge, have three major limitations. They: 1) do not cover the educational domain; 2) typically focus on low-complexity, generic knowledge or broad domains, which do not adequately assess the models' knowledge in specific subjects; and 3) often rely on templates that can bias model predictions. Here, we introduce MALAMUTE, a multilingual, template-free, and highly granular probing dataset comprising expert-written, peer-reviewed probes from 71 university-level textbooks across three languages (English, Spanish, and Polish). MALAMUTE is the first education-based cloze-style dataset. It covers eight domains, each with up to 14 subdomains, further broken down into concepts and concept-based prompts, totaling 33,361 university curriculum concepts and 116,887 prompts. MALAMUTE's fine granularity, educational focus, and inclusion of both sentence-level and paragraph-level prompts make it an ideal tool for evaluating LMs' course-related knowledge. Our evaluation of masked and causal LMs on MALAMUTE shows that despite overall proficiency, they have significant gaps in knowledge when examined closely on specific subjects, hindering their safe use in classrooms and underscoring the need for further development.
☆ RETQA: A Large-Scale Open-Domain Tabular Question Answering Dataset for Real Estate Sector AAAI 2025
The real estate market relies heavily on structured data, such as property details, market trends, and price fluctuations. However, the lack of specialized Tabular Question Answering datasets in this domain limits the development of automated question-answering systems. To fill this gap, we introduce RETQA, the first large-scale open-domain Chinese Tabular Question Answering dataset for Real Estate. RETQA comprises 4,932 tables and 20,762 question-answer pairs across 16 sub-fields within three major domains: property information, real estate company finance information and land auction information. Compared with existing tabular question answering datasets, RETQA poses greater challenges due to three key factors: long-table structures, open-domain retrieval, and multi-domain queries. To tackle these challenges, we propose the SLUTQA framework, which integrates large language models with spoken language understanding tasks to enhance retrieval and answering accuracy. Extensive experiments demonstrate that SLUTQA significantly improves the performance of large language models on RETQA by in-context learning. RETQA and SLUTQA provide essential resources for advancing tabular question answering research in the real estate domain, addressing critical challenges in open-domain and long-table question-answering. The dataset and code are publicly available at \url{https://github.com/jensen-w/RETQA}.
comment: This paper is accepted by AAAI 2025
☆ AMuSeD: An Attentive Deep Neural Network for Multimodal Sarcasm Detection Incorporating Bi-modal Data Augmentation
Detecting sarcasm effectively requires a nuanced understanding of context, including vocal tones and facial expressions. The progression towards multimodal computational methods in sarcasm detection, however, faces challenges due to the scarcity of data. To address this, we present AMuSeD (Attentive deep neural network for MUltimodal Sarcasm dEtection incorporating bi-modal Data augmentation). This approach utilizes the Multimodal Sarcasm Detection Dataset (MUStARD) and introduces a two-phase bimodal data augmentation strategy. The first phase involves generating varied text samples through Back Translation from several secondary languages. The second phase involves the refinement of a FastSpeech 2-based speech synthesis system, tailored specifically for sarcasm to retain sarcastic intonations. Alongside a cloud-based Text-to-Speech (TTS) service, this Fine-tuned FastSpeech 2 system produces corresponding audio for the text augmentations. We also investigate various attention mechanisms for effectively merging text and audio data, finding self-attention to be the most efficient for bimodal integration. Our experiments reveal that this combined augmentation and attention approach achieves a significant F1-score of 81.0% in text-audio modalities, surpassing even models that use three modalities from the MUStARD dataset.
comment: This is a preprint version of the paper, submitted and under review at the IEEE Transactions on Affective Computing
☆ HiTZ at VarDial 2025 NorSID: Overcoming Data Scarcity with Language Transfer and Automatic Data Annotation
In this paper we present our submission for the NorSID Shared Task as part of the 2025 VarDial Workshop (Scherrer et al., 2025), consisting of three tasks: Intent Detection, Slot Filling and Dialect Identification, evaluated using data in different dialects of the Norwegian language. For Intent Detection and Slot Filling, we have fine-tuned a multitask model in a cross-lingual setting, to leverage the xSID dataset available in 17 languages. In the case of Dialect Identification, our final submission consists of a model fine-tuned on the provided development set, which has obtained the highest scores within our experiments. Our final results on the test set show that our models do not drop in performance compared to the development set, likely due to the domain-specificity of the dataset and the similar distribution of both subsets. Finally, we also report an in-depth analysis of the provided datasets and their artifacts, as well as other sets of experiments that have been carried out but did not yield the best results. Additionally, we present an analysis on the reasons why some methods have been more successful than others; mainly the impact of the combination of languages and domain-specificity of the training data on the results.
comment: Vardial 2025 NorSID Shared Task
☆ Lost in the Middle, and In-Between: Enhancing Language Models' Ability to Reason Over Long Contexts in Multi-Hop QA
Previous work finds that recent long-context language models fail to make equal use of information in the middle of their inputs, preferring pieces of information located at the tail ends which creates an undue bias in situations where we would like models to be equally capable of using different parts of the input. Thus far, the problem has mainly only been considered in settings with single pieces of critical information, leading us to question what happens when multiple necessary pieces of information are spread out over the inputs. Here, we demonstrate the effects of the "lost in the middle" problem in the multi-hop question answering setting -- in which multiple reasoning "hops" over disconnected documents are required -- and show that performance degrades not only with respect to the distance of information from the edges of the context, but also between pieces of information. Additionally, we experiment with means of alleviating the problem by reducing superfluous document contents through knowledge graph triple extraction and summarization, and prompting models to reason more thoroughly using chain-of-thought prompting.
☆ GAOKAO-Eval: Does high scores truly reflect strong capabilities in LLMs?
Large Language Models (LLMs) are commonly evaluated using human-crafted benchmarks, under the premise that higher scores implicitly reflect stronger human-like performance. However, there is growing concern that LLMs may ``game" these benchmarks due to data leakage, achieving high scores while struggling with tasks simple for humans. To substantively address the problem, we create GAOKAO-Eval, a comprehensive benchmark based on China's National College Entrance Examination (Gaokao), and conduct ``closed-book" evaluations for representative models released prior to Gaokao. Contrary to prevailing consensus, even after addressing data leakage and comprehensiveness, GAOKAO-Eval reveals that high scores still fail to truly reflect human-aligned capabilities. To better understand this mismatch, We introduce the Rasch model from cognitive psychology to analyze LLM scoring patterns and identify two key discrepancies: 1) anomalous consistent performance across various question difficulties, and 2) high variance in performance on questions of similar difficulty. In addition, We identified inconsistent grading of LLM-generated answers among teachers and recurring mistake patterns. we find that the phenomenons are well-grounded in the motivations behind OpenAI o1, and o1's reasoning-as-difficulties can mitigate the mismatch. These results show that GAOKAO-Eval can reveal limitations in LLM capabilities not captured by current benchmarks and highlight the need for more LLM-aligned difficulty analysis.
comment: 10 pages, 13 figures
☆ Unsupervised Named Entity Disambiguation for Low Resource Domains EMNLP-2024
In the ever-evolving landscape of natural language processing and information retrieval, the need for robust and domain-specific entity linking algorithms has become increasingly apparent. It is crucial in a considerable number of fields such as humanities, technical writing and biomedical sciences to enrich texts with semantics and discover more knowledge. The use of Named Entity Disambiguation (NED) in such domains requires handling noisy texts, low resource settings and domain-specific KBs. Existing approaches are mostly inappropriate for such scenarios, as they either depend on training data or are not flexible enough to work with domain-specific KBs. Thus in this work, we present an unsupervised approach leveraging the concept of Group Steiner Trees (GST), which can identify the most relevant candidates for entity disambiguation using the contextual similarities across candidate entities for all the mentions present in a document. We outperform the state-of-the-art unsupervised methods by more than 40\% (in avg.) in terms of Precision@1 across various domain-specific datasets.
comment: Accepted in EMNLP-2024
☆ Automated Collection of Evaluation Dataset for Semantic Search in Low-Resource Domain Language COLING 2025
Domain-specific languages that use a lot of specific terminology often fall into the category of low-resource languages. Collecting test datasets in a narrow domain is time-consuming and requires skilled human resources with domain knowledge and training for the annotation task. This study addresses the challenge of automated collecting test datasets to evaluate semantic search in low-resource domain-specific German language of the process industry. Our approach proposes an end-to-end annotation pipeline for automated query generation to the score reassessment of query-document pairs. To overcome the lack of text encoders trained in the German chemistry domain, we explore a principle of an ensemble of "weak" text encoders trained on common knowledge datasets. We combine individual relevance scores from diverse models to retrieve document candidates and relevance scores generated by an LLM, aiming to achieve consensus on query-document alignment. Evaluation results demonstrate that the ensemble method significantly improves alignment with human-assigned relevance scores, outperforming individual models in both inter-coder agreement and accuracy metrics. These findings suggest that ensemble learning can effectively adapt semantic search systems for specialized, low-resource languages, offering a practical solution to resource limitations in domain-specific contexts.
comment: accepted in the First Workshop on Language Models for Low-Resource Languages (LoResLM) co-located with the 31st International Conference on Computational Linguistics (COLING 2025)
☆ The role of inhibitory control in garden-path sentence processing: A Chinese-English bilingual perspective
In reading garden-path sentences, people must resolve competing interpretations, though initial misinterpretations can linger despite reanalysis. This study examines the role of inhibitory control (IC) in managing these misinterpretations among Chinese-English bilinguals. Using self-paced reading tasks, we investigated how IC influences recovery from garden-path sentences in Chinese (L1) and its interaction with language proficiency during English (L2) processing. Results indicate that IC does not affect garden-path recovery in Chinese, suggesting reliance on semantic context may reduce the need for IC. In contrast, findings for English L2 learners reveal a complex relationship between language proficiency and IC: Participants with low L2 proficiency but high IC showed lingering misinterpretations, while those with high proficiency exhibited none. These results support and extend the Model of Cognitive Control (Ness et al., 2023). Moreover, our comparison of three Stroop task versions identifies L1 colour-word Stroop task as the preferred measure of IC in bilingual research.
☆ A Comparative Study of LLMs, NMT Models, and Their Combination in Persian-English Idiom Translation
Large language models (LLMs) have shown superior capabilities in translating figurative language compared to neural machine translation (NMT) systems. However, the impact of different prompting methods and LLM-NMT combinations on idiom translation has yet to be thoroughly investigated. This paper introduces two parallel datasets of sentences containing idiomatic expressions for Persian$\rightarrow$English and English$\rightarrow$Persian translations, with Persian idioms sampled from our PersianIdioms resource, a collection of 2,200 idioms and their meanings. Using these datasets, we evaluate various open- and closed-source LLMs, NMT models, and their combinations. Translation quality is assessed through idiom translation accuracy and fluency. We also find that automatic evaluation methods like LLM-as-a-judge, BLEU and BERTScore are effective for comparing different aspects of model performance. Our experiments reveal that Claude-3.5-Sonnet delivers outstanding results in both translation directions. For English$\rightarrow$Persian, combining weaker LLMs with Google Translate improves results, while Persian$\rightarrow$English translations benefit from single prompts for simpler models and complex prompts for advanced ones.
☆ Small Language Model as Data Prospector for Large Language Model
The quality of instruction data directly affects the performance of fine-tuned Large Language Models (LLMs). Previously, \cite{li2023one} proposed \texttt{NUGGETS}, which identifies and selects high-quality quality data from a large dataset by identifying those individual instruction examples that can significantly improve the performance of different tasks after being learnt as one-shot instances. In this work, we propose \texttt{SuperNUGGETS}, an improved variant of \texttt{NUGGETS} optimised for efficiency and performance. Our \texttt{SuperNUGGETS} uses a small language model (SLM) instead of a large language model (LLM) to filter the data for outstanding one-shot instances and refines the predefined set of tests. The experimental results show that the performance of \texttt{SuperNUGGETS} only decreases by 1-2% compared to \texttt{NUGGETS}, but the efficiency can be increased by a factor of 58. Compared to the original \texttt{NUGGETS}, our \texttt{SuperNUGGETS} has a higher utility value due to the significantly lower resource consumption.
☆ Romanized to Native Malayalam Script Transliteration Using an Encoder-Decoder Framework
In this work, we present the development of a reverse transliteration model to convert romanized Malayalam to native script using an encoder-decoder framework built with attention-based bidirectional Long Short Term Memory (Bi-LSTM) architecture. To train the model, we have used curated and combined collection of 4.3 million transliteration pairs derived from publicly available Indic language translitertion datasets, Dakshina and Aksharantar. We evaluated the model on two different test dataset provided by IndoNLP-2025-Shared-Task that contain, (1) General typing patterns and (2) Adhoc typing patterns, respectively. On the Test Set-1, we obtained a character error rate (CER) of 7.4%. However upon Test Set-2, with adhoc typing patterns, where most vowel indicators are missing, our model gave a CER of 22.7%.
comment: 5 pages
☆ Enhancing Nursing and Elderly Care with Large Language Models: An AI-Driven Framework
This paper explores the application of large language models (LLMs) in nursing and elderly care, focusing on AI-driven patient monitoring and interaction. We introduce a novel Chinese nursing dataset and implement incremental pre-training (IPT) and supervised fine-tuning (SFT) techniques to enhance LLM performance in specialized tasks. Using LangChain, we develop a dynamic nursing assistant capable of real-time care and personalized interventions. Experimental results demonstrate significant improvements, paving the way for AI-driven solutions to meet the growing demands of healthcare in aging populations.
☆ Simulating Hard Attention Using Soft Attention
We study conditions under which transformers using soft attention can simulate hard attention, that is, effectively focus all attention on a subset of positions. First, we examine several variants of linear temporal logic, whose formulas have been previously been shown to be computable using hard attention transformers. We demonstrate how soft attention transformers can compute formulas of these logics using unbounded positional embeddings or temperature scaling. Second, we demonstrate how temperature scaling allows softmax transformers to simulate a large subclass of average-hard attention transformers, those that have what we call the uniform-tieless property.
☆ Low-Resource Fast Text Classification Based on Intra-Class and Inter-Class Distance Calculation
In recent years, text classification methods based on neural networks and pre-trained models have gained increasing attention and demonstrated excellent performance. However, these methods still have some limitations in practical applications: (1) They typically focus only on the matching similarity between sentences. However, there exists implicit high-value information both within sentences of the same class and across different classes, which is very crucial for classification tasks. (2) Existing methods such as pre-trained language models and graph-based approaches often consume substantial memory for training and text-graph construction. (3) Although some low-resource methods can achieve good performance, they often suffer from excessively long processing times. To address these challenges, we propose a low-resource and fast text classification model called LFTC. Our approach begins by constructing a compressor list for each class to fully mine the regularity information within intra-class data. We then remove redundant information irrelevant to the target classification to reduce processing time. Finally, we compute the similarity distance between text pairs for classification. We evaluate LFTC on 9 publicly available benchmark datasets, and the results demonstrate significant improvements in performance and processing time, especially under limited computational and data resources, highlighting its superior advantages.
☆ Enhancing the Reasoning Capabilities of Small Language Models via Solution Guidance Fine-Tuning COLING 2025
Large language models (LLMs) have demonstrated remarkable performance across a wide range of tasks. Advances in prompt engineering and fine-tuning techniques have further enhanced their ability to address complex reasoning challenges. However, these advanced capabilities are often exclusive to models exceeding 100 billion parameters. Although Chain-of-Thought (CoT) fine-tuning methods have been explored for smaller models (under 10 billion parameters), they typically depend on extensive CoT training data, which can introduce inconsistencies and limit effectiveness in low-data settings. To overcome these limitations, this paper introduce a new reasoning strategy Solution Guidance (SG) and a plug-and-play training paradigm Solution-Guidance Fine-Tuning (SGFT) for enhancing the reasoning capabilities of small language models. SG focuses on problem understanding and decomposition at the semantic and logical levels, rather than specific computations, which can effectively improve the SLMs' generalization and reasoning abilities. With only a small amount of SG training data, SGFT can fine-tune a SLM to produce accurate problem-solving guidances, which can then be flexibly fed to any SLM as prompts, enabling it to generate correct answers directly. Experimental results demonstrate that our method significantly improves the performance of SLMs on various reasoning tasks, enhancing both their practicality and efficiency within resource-constrained environments.
comment: 11 pages, 4 figures, to be published in The 31st International Conference on Computational Linguistics (COLING 2025)
☆ Analyzing Fairness of Computer Vision and Natural Language Processing Models
Machine learning (ML) algorithms play a crucial role in decision making across diverse fields such as healthcare, finance, education, and law enforcement. Despite their widespread adoption, these systems raise ethical and social concerns due to potential biases and fairness issues. This study focuses on evaluating and improving the fairness of Computer Vision and Natural Language Processing (NLP) models applied to unstructured datasets, emphasizing how biased predictions can reinforce existing systemic inequalities. A publicly available dataset from Kaggle was utilized to simulate a practical scenario for examining fairness in ML workflows. To address and mitigate biases, the study employed two leading fairness libraries: Fairlearn by Microsoft, and AIF360 by IBM. These tools offer comprehensive frameworks for fairness analysis, including metrics evaluation, result visualization, and bias mitigation techniques. The research aims to measure bias levels in ML models, compare the effectiveness of these fairness libraries, and provide actionable recommendations for practitioners. The results demonstrate that each library possesses distinct strengths and limitations in evaluating and mitigating fairness. By systematically analyzing these tools, the study contributes valuable insights to the growing field of ML fairness, offering practical guidance for integrating fairness solutions into real world applications. This research underscores the importance of building more equitable and responsible machine learning systems.
comment: 16 pages, 1 table, 4 figures
☆ Benchmarking Table Comprehension In The Wild
Large Language Models (LLMs), while being increasingly dominant on a myriad of knowledge-intensive activities, have only had limited success understanding lengthy table-text mixtures, such as academic papers and financial reports. Recent advances of long-context LLMs have opened up new possibilities for this field. Nonetheless, we identify two roadblocks: (1) Prior benchmarks of table question answering (TableQA) have focused on isolated tables without context, making it hard to evaluate models in real-world scenarios. (2) Prior benchmarks have focused on some narrow skill sets of table comprehension such as table recognition, data manipulation/calculation, table summarization etc., while a skilled human employs those skills collectively. In this work, we introduce TableQuest, a new benchmark designed to evaluate the holistic table comprehension capabilities of LLMs in the natural table-rich context of financial reports. We employ a rigorous data processing and filtering procedure to ensure that the question-answer pairs are logical, reasonable, and diverse. We experiment with 7 state-of-the-art models, and find that despite reasonable accuracy in locating facts, they often falter when required to execute more sophisticated reasoning or multi-step calculations. We conclude with a qualitative study of the failure modes and discuss the challenges of constructing a challenging benchmark. We make the evaluation data, judging procedure and results of this study publicly available to facilitate research in this field.
comment: Accepted at TRL Workshop@Neurips 2024. Link to data https://github.com/boson-ai/Table_eval_public
☆ On the Limit of Language Models as Planning Formalizers
Large Language Models have been shown to fail to create executable and verifiable plans in grounded environments. An emerging line of work shows success in using LLM as a formalizer to generate a formal representation (e.g., PDDL) of the planning domain, which can be deterministically solved to find a plan. We systematically evaluate this methodology while bridging some major gaps. While previous work only generates a partial PDDL representation given templated and thus unrealistic environment descriptions, we generate the complete representation given descriptions of various naturalness levels. Among an array of observations critical to improve LLMs' formal planning ability, we note that large enough models can effectively formalize descriptions as PDDL, outperforming those directly generating plans, while being robust to lexical perturbation. As the descriptions become more natural-sounding, we observe a decrease in performance and provide detailed error analysis.
☆ Byte Latent Transformer: Patches Scale Better Than Tokens
We introduce the Byte Latent Transformer (BLT), a new byte-level LLM architecture that, for the first time, matches tokenization-based LLM performance at scale with significant improvements in inference efficiency and robustness. BLT encodes bytes into dynamically sized patches, which serve as the primary units of computation. Patches are segmented based on the entropy of the next byte, allocating more compute and model capacity where increased data complexity demands it. We present the first FLOP controlled scaling study of byte-level models up to 8B parameters and 4T training bytes. Our results demonstrate the feasibility of scaling models trained on raw bytes without a fixed vocabulary. Both training and inference efficiency improve due to dynamically selecting long patches when data is predictable, along with qualitative improvements on reasoning and long tail generalization. Overall, for fixed inference costs, BLT shows significantly better scaling than tokenization-based models, by simultaneously growing both patch and model size.
☆ Human-Like Embodied AI Interviewer: Employing Android ERICA in Real International Conference COLING 2025
This paper introduces the human-like embodied AI interviewer which integrates android robots equipped with advanced conversational capabilities, including attentive listening, conversational repairs, and user fluency adaptation. Moreover, it can analyze and present results post-interview. We conducted a real-world case study at SIGDIAL 2024 with 42 participants, of whom 69% reported positive experiences. This study demonstrated the system's effectiveness in conducting interviews just like a human and marked the first employment of such a system at an international conference. The demonstration video is available at https://youtu.be/jCuw9g99KuE.
comment: This paper has been accepted for demonstration presentation at International Conference on Computational Linguistics (COLING 2025)
☆ Financial Sentiment Analysis: Leveraging Actual and Synthetic Data for Supervised Fine-tuning
The Efficient Market Hypothesis (EMH) highlights the essence of financial news in stock price movement. Financial news comes in the form of corporate announcements, news titles, and other forms of digital text. The generation of insights from financial news can be done with sentiment analysis. General-purpose language models are too general for sentiment analysis in finance. Curated labeled data for fine-tuning general-purpose language models are scare, and existing fine-tuned models for sentiment analysis in finance do not capture the maximum context width. We hypothesize that using actual and synthetic data can improve performance. We introduce BertNSP-finance to concatenate shorter financial sentences into longer financial sentences, and finbert-lc to determine sentiment from digital text. The results show improved performance on the accuracy and the f1 score for the financial phrasebank data with $50\%$ and $100\%$ agreement levels.
☆ Low-Rank Adaptation with Task-Relevant Feature Enhancement for Fine-tuning Language Models AAAI 2025
Fine-tuning pre-trained large language models in a parameter-efficient manner is widely studied for its effectiveness and efficiency. LoRA is one of the most widely used methods, which assumes that the optimization process is essentially low dimensional. Although LoRA has demonstrated commendable performance, there remains a significant performance gap between LoRA and full fine-tuning when learning new tasks. In this work, we propose Low-Rank Adaptation with Task-Relevant Feature Enhancement(LoRATRF) for enhancing task-relevant features from the perspective of editing neural network representations. To prioritize task-relevant features, a task-aware filter that selectively extracts valuable knowledge from hidden representations for the target or current task is designed. As the experiments on a vareity of datasets including NLU, commonsense reasoning and mathematical reasoning tasks demonstrates, our method reduces 33.71% parameters and achieves better performance on a variety of datasets in comparison with SOTA low-rank methods.
comment: 6 Pages, 3 figures accepted by AAAI 2025 CoLoRAI - Connecting Low-Rank Representations in AI Workshop
☆ MERaLiON-AudioLLM: Technical Report
We introduce MERaLiON-AudioLLM (Multimodal Empathetic Reasoning and Learning in One Network), the first speech-text model tailored for Singapore's multilingual and multicultural landscape. Developed under the National Large Language Models Funding Initiative, Singapore, MERaLiON-AudioLLM integrates advanced speech and text processing to address the diverse linguistic nuances of local accents and dialects, enhancing accessibility and usability in complex, multilingual environments. Our results demonstrate improvements in both speech recognition and task-specific understanding, positioning MERaLiON-AudioLLM as a pioneering solution for region specific AI applications. We envision this release to set a precedent for future models designed to address localised linguistic and cultural contexts in a global framework.
☆ Enhancing Multimodal Large Language Models Complex Reason via Similarity Computation
Multimodal large language models have experienced rapid growth, and numerous different models have emerged. The interpretability of LVLMs remains an under-explored area. Especially when faced with more complex tasks such as chain-of-thought reasoning, its internal mechanisms still resemble a black box that is difficult to decipher. By studying the interaction and information flow between images and text, we noticed that in models such as LLaVA1.5, image tokens that are semantically related to text are more likely to have information flow convergence in the LLM decoding layer, and these image tokens receive higher attention scores. However, those image tokens that are less relevant to the text do not have information flow convergence, and they only get very small attention scores. To efficiently utilize the image information, we propose a new image token reduction method, Simignore, which aims to improve the complex reasoning ability of LVLMs by computing the similarity between image and text embeddings and ignoring image tokens that are irrelevant and unimportant to the text. Through extensive experiments, we demonstrate the effectiveness of our method for complex reasoning tasks. The paper's source code can be accessed from \url{https://github.com/FanshuoZeng/Simignore}.
☆ ScaleOT: Privacy-utility-scalable Offsite-tuning with Dynamic LayerReplace and Selective Rank Compression AAAI2025
Offsite-tuning is a privacy-preserving method for tuning large language models (LLMs) by sharing a lossy compressed emulator from the LLM owners with data owners for downstream task tuning. This approach protects the privacy of both the model and data owners. However, current offsite tuning methods often suffer from adaptation degradation, high computational costs, and limited protection strength due to uniformly dropping LLM layers or relying on expensive knowledge distillation. To address these issues, we propose ScaleOT, a novel privacy-utility-scalable offsite-tuning framework that effectively balances privacy and utility. ScaleOT introduces a novel layerwise lossy compression algorithm that uses reinforcement learning to obtain the importance of each layer. It employs lightweight networks, termed harmonizers, to replace the raw LLM layers. By combining important original LLM layers and harmonizers in different ratios, ScaleOT generates emulators tailored for optimal performance with various model scales for enhanced privacy protection. Additionally, we present a rank reduction method to further compress the original LLM layers, significantly enhancing privacy with negligible impact on utility. Comprehensive experiments show that ScaleOT can achieve nearly lossless offsite tuning performance compared with full fine-tuning while obtaining better model privacy.
comment: accepted by AAAI2025
LLM Distillation for Efficient Few-Shot Multiple Choice Question Answering
Multiple Choice Question Answering (MCQA) is an important problem with numerous real-world applications, such as medicine, law, and education. The high cost of building MCQA datasets makes few-shot learning pivotal in this domain. While Large Language Models (LLMs) can enable few-shot learning, their direct application in real-world scenarios is often hindered by their high computational cost. To address this challenge, we propose a simple yet effective approach that uses LLMs for data generation and scoring. Our approach utilizes LLMs to create MCQA data which contains questions and choices, and to assign probability scores to the generated choices. We then use the generated data and LLM-assigned scores to finetune a smaller and more efficient encoder-only model, DeBERTa-v3-base by leveraging distillation loss. Extensive experiments on the Massive Multitask Language Understanding (MMLU) benchmark demonstrate that our method improves accuracy from 28.9% to 39.3%, representing a gain of over 10% compared to a baseline finetuned directly on 5-shot examples. This shows the effectiveness of LLM-driven data generation and knowledge distillation for few-shot MCQA.
☆ AutoPatent: A Multi-Agent Framework for Automatic Patent Generation
As the capabilities of Large Language Models (LLMs) continue to advance, the field of patent processing has garnered increased attention within the natural language processing community. However, the majority of research has been concentrated on classification tasks, such as patent categorization and examination, or on short text generation tasks like patent summarization and patent quizzes. In this paper, we introduce a novel and practical task known as Draft2Patent, along with its corresponding D2P benchmark, which challenges LLMs to generate full-length patents averaging 17K tokens based on initial drafts. Patents present a significant challenge to LLMs due to their specialized nature, standardized terminology, and extensive length. We propose a multi-agent framework called AutoPatent which leverages the LLM-based planner agent, writer agents, and examiner agent with PGTree and RRAG to generate lengthy, intricate, and high-quality complete patent documents. The experimental results demonstrate that our AutoPatent framework significantly enhances the ability to generate comprehensive patents across various LLMs. Furthermore, we have discovered that patents generated solely with the AutoPatent framework based on the Qwen2.5-7B model outperform those produced by larger and more powerful LLMs, such as GPT-4o, Qwen2.5-72B, and LLAMA3.1-70B, in both objective metrics and human evaluations. We will make the data and code available upon acceptance at \url{https://github.com/QiYao-Wang/AutoPatent}.
comment: 19 pages, 7 figures
☆ Semi-IIN: Semi-supervised Intra-inter modal Interaction Learning Network for Multimodal Sentiment Analysis
Despite multimodal sentiment analysis being a fertile research ground that merits further investigation, current approaches take up high annotation cost and suffer from label ambiguity, non-amicable to high-quality labeled data acquisition. Furthermore, choosing the right interactions is essential because the significance of intra- or inter-modal interactions can differ among various samples. To this end, we propose Semi-IIN, a Semi-supervised Intra-inter modal Interaction learning Network for multimodal sentiment analysis. Semi-IIN integrates masked attention and gating mechanisms, enabling effective dynamic selection after independently capturing intra- and inter-modal interactive information. Combined with the self-training approach, Semi-IIN fully utilizes the knowledge learned from unlabeled data. Experimental results on two public datasets, MOSI and MOSEI, demonstrate the effectiveness of Semi-IIN, establishing a new state-of-the-art on several metrics. Code is available at https://github.com/flow-ljh/Semi-IIN.
♻ ☆ DroidSpeak: KV Cache Sharing for Efficient Multi-LLM Serving
Large Language Models (LLMs) are increasingly employed in complex workflows, where different LLMs and fine-tuned variants collaboratively address complex tasks. However, these systems face significant inefficiencies due to redundant context processing of the shared context. We propose DroidSpeak, a framework that optimizes context sharing between fine-tuned LLMs derived from the same foundational model. DroidSpeak identifies critical layers in the KV cache and selectively recomputes them, enabling effective reuse of intermediate data while maintaining high accuracy. Our approach balances computational efficiency and task fidelity, significantly reducing inference latency and throughput bottlenecks. Experiments on diverse datasets and model pairs demonstrate that DroidSpeak achieves up to 3x higher throughputs and 2.6x faster prefill times with negligible accuracy loss compared to full recomputation.
♻ ☆ Enhancing Temporal Understanding in Audio Question Answering for Large Audio Language Models
The Audio Question Answering (AQA) task includes audio event classification, audio captioning, and open-ended reasoning. Recently, AQA has garnered attention due to the advent of Large Audio Language Models (LALMs). Current literature focuses on constructing LALMs by integrating audio encoders with text-only Large Language Models (LLMs) through a projection module. While LALMs excel in general audio understanding, they are limited in temporal reasoning, which may hinder their commercial applications and on-device deployment. This paper addresses these challenges and limitations in audio temporal reasoning. First, we introduce a data augmentation technique for generating reliable audio temporal questions and answers using an LLM. Second, we perform a further fine-tuning of an existing baseline using curriculum learning strategy to specialize in temporal reasoning without compromising performance on fine-tuned tasks. We demonstrate the performance of our model using state-of-the-art LALMs on public audio benchmark datasets. Third, we implement our AQA model on-device locally and investigate its CPU inference for edge applications.
comment: 9 pages, 6 figures
♻ ☆ NLP Cluster Analysis of Common Core State Standards and NAEP Item Specifications
Camilli (2024) proposed a methodology using natural language processing (NLP) to map the relationship of a set of content standards to item specifications. This study provided evidence that NLP can be used to improve the mapping process. As part of this investigation, the nominal classifications of standards and items specifications were used to examine construct equivalence. In the current paper, we determine the strength of empirical support for the semantic distinctiveness of these classifications, which are known as "domains" for Common Core standards, and "strands" for National Assessment of Educational Progress (NAEP) item specifications. This is accomplished by separate k-means clustering for standards and specifications of their corresponding embedding vectors. We then briefly illustrate an application of these findings.
comment: 10 pages, 5 tables
♻ ☆ Building Better: Avoiding Pitfalls in Developing Language Resources when Data is Scarce
Language is a symbolic capital that affects people's lives in many ways (Bourdieu, 1977, 1991). It is a powerful tool that accounts for identities, cultures, traditions, and societies in general. Hence, data in a given language should be viewed as more than a collection of tokens. Good data collection and labeling practices are key to building more human-centered and socially aware technologies. While there has been a rising interest in mid- to low-resource languages within the NLP community, work in this space has to overcome unique challenges such as data scarcity and access to suitable annotators. In this paper, we collect feedback from those directly involved in and impacted by NLP artefacts for mid- to low-resource languages. We conduct a quantitative and qualitative analysis of the responses and highlight the main issues related to (1) data quality such as linguistic and cultural data suitability; and (2) the ethics of common annotation practices such as the misuse of online community services. Based on these findings, we make several recommendations for the creation of high-quality language artefacts that reflect the cultural milieu of its speakers, while simultaneously respecting the dignity and labor of data workers.
♻ ☆ Linguistic Minimal Pairs Elicit Linguistic Similarity in Large Language Models COLING 2025
We introduce a novel analysis that leverages linguistic minimal pairs to probe the internal linguistic representations of Large Language Models (LLMs). By measuring the similarity between LLM activation differences across minimal pairs, we quantify the and gain insight into the linguistic knowledge captured by LLMs. Our large-scale experiments, spanning 100+ LLMs and 150k minimal pairs in three languages, reveal properties of linguistic similarity from four key aspects: consistency across LLMs, relation to theoretical categorizations, dependency to semantic context, and cross-lingual alignment of relevant phenomena. Our findings suggest that 1) linguistic similarity is significantly influenced by training data exposure, leading to higher cross-LLM agreement in higher-resource languages. 2) Linguistic similarity strongly aligns with fine-grained theoretical linguistic categories but weakly with broader ones. 3) Linguistic similarity shows a weak correlation with semantic similarity, showing its context-dependent nature. 4) LLMs exhibit limited cross-lingual alignment in their understanding of relevant linguistic phenomena. This work demonstrates the potential of minimal pairs as a window into the neural representations of language in LLMs, shedding light on the relationship between LLMs and linguistic theory. Codes and data are available at https://github.com/ChenDelong1999/Linguistic-Similarity
comment: COLING 2025
♻ ☆ Fine Tuning Large Language Models for Medicine: The Role and Importance of Direct Preference Optimization
Large Language Model (LLM) fine tuning is underutilized in the field of medicine. Two of the most common methods of fine tuning are Supervised Fine Tuning (SFT) and Direct Preference Optimization (DPO), but there is little guidance informing users when to use either technique. In this investigation, we compare the performance of SFT and DPO for five common natural language tasks in medicine: Classification with text data, Classification with numeric data, Clinical Reasoning, Summarization, and Clinical Triage. We find that SFT alone is sufficient for Classification with text data, whereas DPO improves performance for the more complex tasks of Clinical Reasoning, Summarization and Clinical Triage. Our results establish the role and importance of DPO fine tuning within medicine, and consequently call attention to current software gaps that prevent widespread deployment of this technique.
♻ ☆ TrustUQA: A Trustful Framework for Unified Structured Data Question Answering AAAI 2025
Natural language question answering (QA) over structured data sources such as tables and knowledge graphs have been widely investigated, especially with Large Language Models (LLMs) in recent years. The main solutions include question to formal query parsing and retrieval-based answer generation. However, current methods of the former often suffer from weak generalization, failing to dealing with multi-types of sources, while the later is limited in trustfulness. In this paper, we propose TrustUQA, a trustful QA framework that can simultaneously support multiple types of structured data in a unified way. To this end, it adopts an LLM-friendly and unified knowledge representation method called Condition Graph(CG), and uses an LLM and demonstration-based two-level method for CG querying. For enhancement, it is also equipped with dynamic demonstration retrieval. We have evaluated TrustUQA with 5 benchmarks covering 3 types of structured data. It outperforms 2 existing unified structured data QA methods. In comparison with the baselines that are specific to one data type, it achieves state-of-the-art on 2 of the datasets. Further more, we have demonstrated the potential of our method for more general QA tasks, QA over mixed structured data and QA across structured data. The code is available at https://github.com/zjukg/TrustUQA.
comment: Accepted by AAAI 2025
♻ ☆ Bridging Sequence-Structure Alignment in RNA Foundation Models AAAI 2025
The alignment between RNA sequences and structures in foundation models (FMs) has yet to be thoroughly investigated. Existing FMs have struggled to establish sequence-structure alignment, hindering the free flow of genomic information between RNA sequences and structures. In this study, we introduce OmniGenome, an RNA FM trained to align RNA sequences with respect to secondary structures based on structure-contextualised modelling. The alignment enables free and bidirectional mappings between sequences and structures by utilising the flexible RNA modelling paradigm that supports versatile input and output modalities, i.e., sequence and/or structure as input/output. We implement RNA design and zero-shot secondary structure prediction as case studies to evaluate the Seq2Str and Str2Seq mapping capacity of OmniGenome. Results on the EternaV2 benchmark show that OmniGenome solved 74% of puzzles, whereas existing FMs only solved up to 3% of the puzzles due to the oversight of sequence-structure alignment. We leverage four comprehensive in-silico genome modelling benchmarks to evaluate performance across a diverse set of genome downstream tasks, where the results show that OmniGenome achieves state-of-the-art performance on RNA and DNA benchmarks, even without any training on DNA genomes.
comment: Accepted by AAAI 2025
♻ ☆ Citation Amnesia: On The Recency Bias of NLP and Other Academic Fields
This study examines the tendency to cite older work across 20 fields of study over 43 years (1980--2023). We put NLP's propensity to cite older work in the context of these 20 other fields to analyze whether NLP shows similar temporal citation patterns to these other fields over time or whether differences can be observed. Our analysis, based on a dataset of approximately 240 million papers, reveals a broader scientific trend: many fields have markedly declined in citing older works (e.g., psychology, computer science). We term this decline a 'citation age recession', analogous to how economists define periods of reduced economic activity. The trend is strongest in NLP and ML research (-12.8% and -5.5% in citation age from previous peaks). Our results suggest that citing more recent works is not directly driven by the growth in publication rates (-3.4% across fields; -5.2% in humanities; -5.5% in formal sciences) -- even when controlling for an increase in the volume of papers. Our findings raise questions about the scientific community's engagement with past literature, particularly for NLP, and the potential consequences of neglecting older but relevant research. The data and a demo showcasing our results are publicly available.
♻ ☆ Searching for Structure: Investigating Emergent Communication with Large Language Models
Human languages have evolved to be structured through repeated language learning and use. These processes introduce biases that operate during language acquisition and shape linguistic systems toward communicative efficiency. In this paper, we investigate whether the same happens if artificial languages are optimised for implicit biases of Large Language Models (LLMs). To this end, we simulate a classical referential game in which LLMs learn and use artificial languages. Our results show that initially unstructured holistic languages are indeed shaped to have some structural properties that allow two LLM agents to communicate successfully. Similar to observations in human experiments, generational transmission increases the learnability of languages, but can at the same time result in non-humanlike degenerate vocabularies. Taken together, this work extends experimental findings, shows that LLMs can be used as tools in simulations of language evolution, and opens possibilities for future human-machine experiments in this field.
♻ ☆ Olympus: A Universal Task Router for Computer Vision Tasks
We introduce Olympus, a new approach that transforms Multimodal Large Language Models (MLLMs) into a unified framework capable of handling a wide array of computer vision tasks. Utilizing a controller MLLM, Olympus delegates over 20 specialized tasks across images, videos, and 3D objects to dedicated modules. This instruction-based routing enables complex workflows through chained actions without the need for training heavy generative models. Olympus easily integrates with existing MLLMs, expanding their capabilities with comparable performance. Experimental results demonstrate that Olympus achieves an average routing accuracy of 94.75% across 20 tasks and precision of 91.82% in chained action scenarios, showcasing its effectiveness as a universal task router that can solve a diverse range of computer vision tasks. Project page: http://yuanze-lin.me/Olympus_page/
comment: Technical Report
♻ ☆ Frequency matters: Modeling irregular morphological patterns in Spanish with Transformers
The present paper evaluates the learning behaviour of a transformer-based neural network with regard to an irregular inflectional paradigm. We apply the paradigm cell filling problem to irregular patterns. We approach this problem using the morphological reinflection task and model it as a character sequence-to-sequence learning problem. The test case under investigation are irregular verbs in Spanish. Besides many regular verbs in Spanish L-shaped verbs the first person singular indicative stem irregularly matches the subjunctive paradigm, while other indicative forms remain unaltered. We examine the role of frequency during learning and compare models under differing input frequency conditions. We train the model on a corpus of Spanish with a realistic distribution of regular and irregular verbs to compare it with models trained on input with augmented distributions of (ir)regular words. We explore how the neural models learn this L-shaped pattern using post-hoc analyses. Our experiments show that, across frequency conditions, the models are surprisingly capable of learning the irregular pattern. Furthermore, our post-hoc analyses reveal the possible sources of errors. All code and data are available at \url{https://anonymous.4open.science/r/modeling_spanish_acl-7567/} under MIT license.
comment: Typos and grammatical corrections
♻ ☆ Neural Text Normalization for Luxembourgish using Real-Life Variation Data
Orthographic variation is very common in Luxembourgish texts due to the absence of a fully-fledged standard variety. Additionally, developing NLP tools for Luxembourgish is a difficult task given the lack of annotated and parallel data, which is exacerbated by ongoing standardization. In this paper, we propose the first sequence-to-sequence normalization models using the ByT5 and mT5 architectures with training data obtained from word-level real-life variation data. We perform a fine-grained, linguistically-motivated evaluation to test byte-based, word-based and pipeline-based models for their strengths and weaknesses in text normalization. We show that our sequence model using real-life variation data is an effective approach for tailor-made normalization in Luxembourgish.
comment: Accepted at VarDial 2025
♻ ☆ Towards Cross-Lingual Audio Abuse Detection in Low-Resource Settings with Few-Shot Learning COLING 2025
Online abusive content detection, particularly in low-resource settings and within the audio modality, remains underexplored. We investigate the potential of pre-trained audio representations for detecting abusive language in low-resource languages, in this case, in Indian languages using Few Shot Learning (FSL). Leveraging powerful representations from models such as Wav2Vec and Whisper, we explore cross-lingual abuse detection using the ADIMA dataset with FSL. Our approach integrates these representations within the Model-Agnostic Meta-Learning (MAML) framework to classify abusive language in 10 languages. We experiment with various shot sizes (50-200) evaluating the impact of limited data on performance. Additionally, a feature visualization study was conducted to better understand model behaviour. This study highlights the generalization ability of pre-trained models in low-resource scenarios and offers valuable insights into detecting abusive language in multilingual contexts.
comment: Accepted as part of the proceedings of COLING 2025
♻ ☆ A Character-Centric Creative Story Generation via Imagination
Creative story generation has long been a goal of NLP research. While existing methodologies have aimed to generate long and coherent stories, they fall significantly short of human capabilities in terms of diversity and character depth. To address this, we introduce a novel story generation framework called CCI (Character-centric Creative story generation via Imagination). CCI features two modules for creative story generation: IG (Image-Guided Imagination) and MW (Multi-Writer model). In the IG module, we utilize a text-to-image model to create visual representations of key story elements, such as characters, backgrounds, and main plots, in a more novel and concrete manner than text-only approaches. The MW module uses these story elements to generate multiple persona-description candidates and selects the best one to insert into the story, thereby enhancing the richness and depth of the narrative. We compared the stories generated by CCI and baseline models through statistical analysis, as well as human and LLM evaluations. The results showed that the IG and MW modules significantly improve various aspects of the stories' creativity. Furthermore, our framework enables interactive multi-modal story generation with users, opening up new possibilities for human-LLM integration in cultural development. Project page : https://www.2024cci.p-e.kr/
♻ ☆ GATEAU: Selecting Influential Sample for Long Context Alignment
Aligning large language models to handle instructions with extremely long contexts has yet to be fully investigated. Previous studies attempt to scale up the available data volume by synthesizing long instruction-following samples, as constructing such a dataset tends to be challenging for annotators. However, a lack of a well-defined strategy for ensuring data quality may introduce low-quality samples and restrict the model performance. Thus, we propose GATEAU, a novel framework to address the unique challenge of long context alignment by identifying the influential samples enriched with long-range dependency relations. Specifically, GATEAU measures the long-range dependencies from two essential aspects: the difficulty of generating target responses due to the long-range dependencies, and the difficulty of understanding long inputs due to such dependencies. Comprehensive experiments indicate that GATEAU effectively identifies influential samples and the model trained on these selected samples exhibits better instruction-following and long-context understanding capabilities.
♻ ☆ Social Bias in Large Language Models For Bangla: An Empirical Study on Gender and Religious Bias COLING 2025
The rapid growth of Large Language Models (LLMs) has put forward the study of biases as a crucial field. It is important to assess the influence of different types of biases embedded in LLMs to ensure fair use in sensitive fields. Although there have been extensive works on bias assessment in English, such efforts are rare and scarce for a major language like Bangla. In this work, we examine two types of social biases in LLM generated outputs for Bangla language. Our main contributions in this work are: (1) bias studies on two different social biases for Bangla, (2) a curated dataset for bias measurement benchmarking and (3) testing two different probing techniques for bias detection in the context of Bangla. This is the first work of such kind involving bias assessment of LLMs for Bangla to the best of our knowledge. All our code and resources are publicly available for the progress of bias related research in Bangla NLP.
comment: Accepted at The First Workshop on Language Models for Low-Resource Languages (LoResLM) at COLING 2025
♻ ☆ TreeEval: Benchmark-Free Evaluation of Large Language Models through Tree Planning
Recently, numerous new benchmarks have been established to evaluate the performance of large language models (LLMs) via either computing a holistic score or employing another LLM as a judge. However, these approaches suffer from data leakage due to the open access of the benchmark and inflexible evaluation process. To address this issue, we introduce $\textbf{TreeEval}$, a benchmark-free evaluation method for LLMs that let a high-performance LLM host an irreproducible evaluation session and essentially avoids the data leakage. Moreover, this LLM performs as an examiner to raise up a series of questions under a topic with a tree planing strategy, which considers the current evaluation status to decide the next question generation and ensures the completeness and efficiency of the evaluation process. We evaluate $6$ models of different parameter sizes, including $7$B, $13$B, and $33$B, and ultimately achieved the highest correlation coefficient with AlpacaEval2.0 using only around $45$ questions. We also conduct more analysis to show the robustness and reliability of TreeEval. Our code can be accessed via the provided https://github.com/Ashura5/TreeEval.
♻ ☆ Clustering Algorithms and RAG Enhancing Semi-Supervised Text Classification with Large LLMs
This paper introduces a novel semi-supervised learning framework specifically designed for text classification tasks, effectively addressing the challenge of vast datasets with limited labeled examples. By integrating multi-level similarity based data augmentation techniques from Retrieval-Augmented Generation (RAG) to Large Language Model (LLM) rewriting and traditional word substitution-we constructed an intelligent augmentation pipeline. This framework innovatively employs the selection of representative landmarks through clustering, which serve as intermediaries in the retrieval and rewriting processes, ensuring that the augmented data maintains a distribution similar to the original dataset. Empirical results show that even in complex text document classification scenarios with over 100 categories, our method achieves state-of-the-art accuracies of 95.41% and 82.43% on the Reuters and Web of Science datasets, respectively. These findings highlight the effectiveness and broad applicability of our semi-supervised learning approach for text classification tasks.
♻ ☆ Strategic Insights in Human and Large Language Model Tactics at Word Guessing Games
At the beginning of 2022, a simplistic word-guessing game took the world by storm and was further adapted to many languages beyond the original English version. In this paper, we examine the strategies of daily word-guessing game players that have evolved during a period of over two years. A survey gathered from 25% of frequent players reveals their strategies and motivations for continuing the daily journey. We also explore the capability of several popular open-access large language model systems and open-source models at comprehending and playing the game in two different languages. Results highlight the struggles of certain models to maintain correct guess length and generate repetitions, as well as hallucinations of non-existent words and inflections.
♻ ☆ Benchmarking LLMs for Mimicking Child-Caregiver Language in Interaction
LLMs can generate human-like dialogues, yet their ability to simulate early child-adult interactions remains largely unexplored. In this paper, we examined how effectively LLMs can capture the distinctive features of child-caregiver language in interaction, using both static and interactive benchmarking methods. We found that state-of-the-art LLMs like Llama 3 and GPT-4o can approximate child-caregiver dialogues at the word and utterance level, but they struggle to reproduce the child and caregiver's discursive patterns, exaggerate alignment, and fail to reach the level of diversity shown by humans. The broader goal of this work is to initiate the development of a comprehensive benchmark for LLMs in child-oriented applications.
♻ ☆ Dynamic Fog Computing for Enhanced LLM Execution in Medical Applications
The ability of large language models (LLMs) to transform, interpret, and comprehend vast quantities of heterogeneous data presents a significant opportunity to enhance data-driven care delivery. However, the sensitive nature of protected health information (PHI) raises valid concerns about data privacy and trust in remote LLM platforms. In addition, the cost associated with cloud-based artificial intelligence (AI) services continues to impede widespread adoption. To address these challenges, we propose a shift in the LLM execution environment from opaque, centralized cloud providers to a decentralized and dynamic fog computing architecture. By executing open-weight LLMs in more trusted environments, such as the user's edge device or a fog layer within a local network, we aim to mitigate the privacy, trust, and financial challenges associated with cloud-based LLMs. We further present SpeziLLM, an open-source framework designed to facilitate rapid and seamless leveraging of different LLM execution layers and lowering barriers to LLM integration in digital health applications. We demonstrate SpeziLLM's broad applicability across six digital health applications, showcasing its versatility in various healthcare settings.
♻ ☆ ViTHSD: Exploiting Hatred by Targets for Hate Speech Detection on Vietnamese Social Media Texts
The growth of social networks makes toxic content spread rapidly. Hate speech detection is a task to help decrease the number of harmful comments. With the diversity in the hate speech created by users, it is necessary to interpret the hate speech besides detecting it. Hence, we propose a methodology to construct a system for targeted hate speech detection from online streaming texts from social media. We first introduce the ViTHSD - a targeted hate speech detection dataset for Vietnamese Social Media Texts. The dataset contains 10K comments, each comment is labeled to specific targets with three levels: clean, offensive, and hate. There are 5 targets in the dataset, and each target is labeled with the corresponding level manually by humans with strict annotation guidelines. The inter-annotator agreement obtained from the dataset is 0.45 by Cohen's Kappa index, which is indicated as a moderate level. Then, we construct a baseline for this task by combining the Bi-GRU-LSTM-CNN with the pre-trained language model to leverage the power of text representation of BERTology. Finally, we suggest a methodology to integrate the baseline model for targeted hate speech detection into the online streaming system for practical application in preventing hateful and offensive content on social media.
comment: Accepted for publication at Journal of Computational Social Science
♻ ☆ Evaluation of Code LLMs on Geospatial Code Generation SP
Software development support tools have been studied for a long time, with recent approaches using Large Language Models (LLMs) for code generation. These models can generate Python code for data science and machine learning applications. LLMs are helpful for software engineers because they increase productivity in daily work. An LLM can also serve as a "mentor" for inexperienced software developers, and be a viable learning support. High-quality code generation with LLMs can also be beneficial in geospatial data science. However, this domain poses different challenges, and code generation LLMs are typically not evaluated on geospatial tasks. Here, we show how we constructed an evaluation benchmark for code generation models, based on a selection of geospatial tasks. We categorised geospatial tasks based on their complexity and required tools. Then, we created a dataset with tasks that test model capabilities in spatial reasoning, spatial data processing, and geospatial tools usage. The dataset consists of specific coding problems that were manually created for high quality. For every problem, we proposed a set of test scenarios that make it possible to automatically check the generated code for correctness. In addition, we tested a selection of existing code generation LLMs for code generation in the geospatial domain. We share our dataset and reproducible evaluation code on a public GitHub repository, arguing that this can serve as an evaluation benchmark for new LLMs in the future. Our dataset will hopefully contribute to the development new models capable of solving geospatial coding tasks with high accuracy. These models will enable the creation of coding assistants tailored for geospatial applications.
comment: 7th ACM SIGSPATIAL International Workshop on AI for Geographic Knowledge Discovery (GeoAI'24)
♻ ☆ CCT-Code: Cross-Consistency Training for Multilingual Clone Detection and Code Search
We consider the well-known and important tasks of clone detection and information retrieval for source code. The most standard setup is to search clones inside the same language code snippets. But it is also useful to find code snippets with identical behaviour in different programming languages. Nevertheless multi- and cross-lingual clone detection has been little studied in literature. We present a novel training procedure, cross-consistency training (CCT) leveraging cross-lingual similarity, that we apply to train language models on source code in various programming languages. We show that this training is effective both for encoder- and decoder-based models. The trained encoder-based CCT-LM model achieves a new state of the art on POJ-104 (monolingual C++ clone detection benchmark) with 96.73\% MAP and AdvTest (monolingual Python code search benchmark) with 47.18\% MRR. The decoder-based CCT-LM model shows comparable performance in these tasks. In addition, we formulate the multi- and cross-lingual clone detection problem and present XCD, a new benchmark dataset produced from CodeForces submissions.
♻ ☆ Learn and Unlearn in Multilingual LLMs
This paper investigates the propagation of harmful information in multilingual large language models (LLMs) and evaluates the efficacy of various unlearning methods. We demonstrate that fake information, regardless of the language it is in, once introduced into these models through training data, can spread across different languages, compromising the integrity and reliability of the generated content. Our findings reveal that standard unlearning techniques, which typically focus on English data, are insufficient in mitigating the spread of harmful content in multilingual contexts and could inadvertently reinforce harmful content across languages. We show that only by addressing harmful responses in both English and the original language of the harmful data can we effectively eliminate generations for all languages. This underscores the critical need for comprehensive unlearning strategies that consider the multilingual nature of modern LLMs to enhance their safety and reliability across diverse linguistic landscapes.
♻ ☆ Bootstrapping Heterogeneous Graph Representation Learning via Large Language Models: A Generalized Approach AAAI 2025
Graph representation learning methods are highly effective in handling complex non-Euclidean data by capturing intricate relationships and features within graph structures. However, traditional methods face challenges when dealing with heterogeneous graphs that contain various types of nodes and edges due to the diverse sources and complex nature of the data. Existing Heterogeneous Graph Neural Networks (HGNNs) have shown promising results but require prior knowledge of node and edge types and unified node feature formats, which limits their applicability. Recent advancements in graph representation learning using Large Language Models (LLMs) offer new solutions by integrating LLMs' data processing capabilities, enabling the alignment of various graph representations. Nevertheless, these methods often overlook heterogeneous graph data and require extensive preprocessing. To address these limitations, we propose a novel method that leverages the strengths of both LLM and GNN, allowing for the processing of graph data with any format and type of nodes and edges without the need for type information or special preprocessing. Our method employs LLM to automatically summarize and classify different data formats and types, aligns node features, and uses a specialized GNN for targeted learning, thus obtaining effective graph representations for downstream tasks. Theoretical analysis and experimental validation have demonstrated the effectiveness of our method.
comment: Accepted by AAAI 2025
♻ ☆ First Train to Generate, then Generate to Train: UnitedSynT5 for Few-Shot NLI
Natural Language Inference (NLI) tasks require identifying the relationship between sentence pairs, typically classified as entailment, contradiction, or neutrality. While the current state-of-the-art (SOTA) model, Entailment Few-Shot Learning (EFL), achieves a 93.1% accuracy on the Stanford Natural Language Inference (SNLI) dataset, further advancements are constrained by the dataset's limitations. To address this, we propose a novel approach leveraging synthetic data augmentation to enhance dataset diversity and complexity. We present UnitedSynT5, an advanced extension of EFL that leverages a T5-based generator to synthesize additional premise-hypothesis pairs, which are rigorously cleaned and integrated into the training data. These augmented examples are processed within the EFL framework, embedding labels directly into hypotheses for consistency. We train a GTR-T5-XL model on this expanded dataset, achieving a new benchmark of 94.7% accuracy on the SNLI dataset, 94.0% accuracy on the E-SNLI dataset, and 92.6% accuracy on the MultiNLI dataset, surpassing the previous SOTA models. This research demonstrates the potential of synthetic data augmentation in improving NLI models, offering a path forward for further advancements in natural language understanding tasks.
comment: 14 pages
♻ ☆ Role-playing Prompt Framework: Generation and Evaluation
Large language models (LLMs) exhibit impressive proficiency in natural language generation, understanding user instructions, and emulating human-like language use, which has led to significant interest in their application to role-playing scenarios. However, the manual collection of role-specific script data and the evaluation of model performance are resource-intensive processes. This paper introduces a prompt-based framework designed to leverage GPT's capabilities for the generation of role-playing dialogue datasets and the evaluation of role-playing performance. To validate the effectiveness of the GPT-based generation and evaluation, we further incorporate the recall-oriented Rouge-L metric, providing an additional quantitative measure of performance.
♻ ☆ Towards Efficient Methods in Medical Question Answering using Knowledge Graph Embeddings
In Natural Language Processing (NLP), Machine Reading Comprehension (MRC) is the task of answering a question based on a given context. To handle questions in the medical domain, modern language models such as BioBERT, SciBERT and even ChatGPT are trained on vast amounts of in-domain medical corpora. However, in-domain pre-training is expensive in terms of time and resources. In this paper, we propose a resource-efficient approach for injecting domain knowledge into a model without relying on such domain-specific pre-training. Knowledge graphs are powerful resources for accessing medical information. Building on existing work, we introduce a method using Multi-Layer Perceptrons (MLPs) for aligning and integrating embeddings extracted from medical knowledge graphs with the embedding spaces of pre-trained language models (LMs). The aligned embeddings are fused with open-domain LMs BERT and RoBERTa that are fine-tuned for two MRC tasks, span detection (COVID-QA) and multiple-choice questions (PubMedQA). We compare our method to prior techniques that rely on a vocabulary overlap for embedding alignment and show how our method circumvents this requirement to deliver better performance. On both datasets, our method allows BERT/RoBERTa to either perform on par (occasionally exceeding) with stronger domain-specific models or show improvements in general over prior techniques. With the proposed approach, we signal an alternative method to in-domain pre-training to achieve domain proficiency. Our code is available here.
comment: Accepted to the MABM workshop at IEEE BIBM 2024
♻ ☆ Pretraining Vision-Language Model for Difference Visual Question Answering in Longitudinal Chest X-rays
Difference visual question answering (diff-VQA) is a challenging task that requires answering complex questions based on differences between a pair of images. This task is particularly important in reading chest X-ray images because radiologists often compare multiple images of the same patient taken at different times to track disease progression and changes in its severity in their clinical practice. However, previous works focused on designing specific network architectures for the diff-VQA task, missing opportunities to enhance the model's performance using a pretrained vision-language model (VLM). Here, we introduce a novel VLM called PLURAL, which is pretrained on natural and longitudinal chest X-ray data for the diff-VQA task. The model is developed using a step-by-step approach, starting with being pretrained on natural images and texts, followed by being trained using longitudinal chest X-ray data. The longitudinal data consist of pairs of X-ray images, along with question-answer sets and radiologist's reports that describe the changes in lung abnormalities and diseases over time. Our experimental results show that the PLURAL model outperforms state-of-the-art methods not only in diff-VQA for longitudinal X-rays but also in conventional VQA for a single X-ray image. Through extensive experiments, we demonstrate the effectiveness of the proposed VLM architecture and pretraining method in improving the model's performance.
♻ ☆ ReFT: Reasoning with Reinforced Fine-Tuning ACL 2024
One way to enhance the reasoning capability of Large Language Models (LLMs) is to conduct Supervised Fine-Tuning (SFT) using Chain-of-Thought (CoT) annotations. This approach does not show sufficiently strong generalization ability, however, because the training only relies on the given CoT data. In math problem-solving, for example, there is usually only one annotated reasoning path for each question in the training data. Intuitively, it would be better for the algorithm to learn from multiple annotated reasoning paths given a question. To address this issue, we propose a simple yet effective approach called Reinforced Fine-Tuning (ReFT) to enhance the generalizability of learning LLMs for reasoning, with math problem-solving as an example. ReFT first warmups the model with SFT, and then employs on-line reinforcement learning, specifically the PPO algorithm in this paper, to further fine-tune the model, where an abundance of reasoning paths are automatically sampled given the question and the rewards are naturally derived from the ground-truth answers. Extensive experiments on GSM8K, MathQA, and SVAMP datasets show that ReFT significantly outperforms SFT, and the performance can be potentially further boosted by combining inference-time strategies such as majority voting and re-ranking. Note that ReFT obtains the improvement by learning from the same training questions as SFT, without relying on extra or augmented training questions. This indicates a superior generalization ability for ReFT.
comment: ACL 2024 main conference; adjust with reviewer comments; 13 pages
♻ ☆ Political Actor Agent: Simulating Legislative System for Roll Call Votes Prediction with Large Language Models AAAI 2025
Predicting roll call votes through modeling political actors has emerged as a focus in quantitative political science and computer science. Widely used embedding-based methods generate vectors for legislators from diverse data sets to predict legislative behaviors. However, these methods often contend with challenges such as the need for manually predefined features, reliance on extensive training data, and a lack of interpretability. Achieving more interpretable predictions under flexible conditions remains an unresolved issue. This paper introduces the Political Actor Agent (PAA), a novel agent-based framework that utilizes Large Language Models to overcome these limitations. By employing role-playing architectures and simulating legislative system, PAA provides a scalable and interpretable paradigm for predicting roll-call votes. Our approach not only enhances the accuracy of predictions but also offers multi-view, human-understandable decision reasoning, providing new insights into political actor behaviors. We conducted comprehensive experiments using voting records from the 117-118th U.S. House of Representatives, validating the superior performance and interpretability of PAA. This study not only demonstrates PAA's effectiveness but also its potential in political science research.
comment: Accepted at AAAI 2025
♻ ☆ Improving Factuality in Large Language Models via Decoding-Time Hallucinatory and Truthful Comparators AAAI 2025
Despite their remarkable capabilities, Large Language Models (LLMs) are prone to generate responses that contradict verifiable facts, i.e., unfaithful hallucination content. Existing efforts generally focus on optimizing model parameters or editing semantic representations, which compromise the internal factual knowledge of target LLMs. In addition, hallucinations typically exhibit multifaceted patterns in downstream tasks, limiting the model's holistic performance across tasks. In this paper, we propose a Comparator-driven Decoding-Time (CDT) framework to alleviate the response hallucination. Firstly, we construct hallucinatory and truthful comparators with multi-task fine-tuning samples. In this case, we present an instruction prototype-guided mixture of experts strategy to enhance the ability of the corresponding comparators to capture different hallucination or truthfulness patterns in distinct task instructions. CDT constrains next-token predictions to factuality-robust distributions by contrasting the logit differences between the target LLMs and these comparators. Systematic experiments on multiple downstream tasks show that our framework can significantly improve the model performance and response factuality.
comment: Accepted by AAAI 2025
♻ ☆ AutoDCWorkflow: LLM-based Data Cleaning Workflow Auto-Generation and Benchmark
We investigate the reasoning capabilities of large language models (LLMs) for automatically generating data-cleaning workflows. To evaluate LLMs' ability to complete data-cleaning tasks, we implemented a pipeline for LLM-based Auto Data Cleaning Workflow (AutoDCWorkflow), prompting LLMs on data cleaning operations to repair three types of data quality issues: duplicates, missing values, and inconsistent data formats. Given a dirty table and a purpose (expressed as a query), this pipeline generates a minimal, clean table sufficient to address the purpose and the data cleaning workflow used to produce the table. The planning process involves three main LLM-driven components: (1) Select Target Columns: Identifies a set of target columns related to the purpose. (2) Inspect Column Quality: Assesses the data quality for each target column and generates a Data Quality Report as operation objectives. (3) Generate Operation & Arguments: Predicts the next operation and arguments based on the data quality report results. Additionally, we propose a data cleaning benchmark to evaluate the capability of LLM agents to automatically generate workflows that address data cleaning purposes of varying difficulty levels. The benchmark comprises the annotated datasets as a collection of purpose, raw table, clean table, data cleaning workflow, and answer set. In our experiments, we evaluated three LLMs that auto-generate purpose-driven data cleaning workflows. The results indicate that LLMs perform well in planning and generating data-cleaning workflows without the need for fine-tuning.
♻ ☆ Too Late to Train, Too Early To Use? A Study on Necessity and Viability of Low-Resource Bengali LLMs
Each new generation of English-oriented Large Language Models (LLMs) exhibits enhanced cross-lingual transfer capabilities and significantly outperforms older LLMs on low-resource languages. This prompts the question: Is there a need for LLMs dedicated to a particular low-resource language? We aim to explore this question for Bengali, a low-to-moderate resource Indo-Aryan language native to the Bengal region of South Asia. We compare the performance of open-weight and closed-source LLMs such as LLaMA-3 and GPT-4 against fine-tuned encoder-decoder models across a diverse set of Bengali downstream tasks, including translation, summarization, paraphrasing, question-answering, and natural language inference. Our findings reveal that while LLMs generally excel in reasoning tasks, their performance in tasks requiring Bengali script generation is inconsistent. Key challenges include inefficient tokenization of Bengali script by existing LLMs, leading to increased computational costs and potential performance degradation. Additionally, we highlight biases in machine-translated datasets commonly used for Bengali NLP tasks. We conclude that there is a significant need for a Bengali-oriented LLM, but the field currently lacks the high-quality pretraining and instruction-tuning datasets necessary to develop a highly effective model.
♻ ☆ Recording for Eyes, Not Echoing to Ears: Contextualized Spoken-to-Written Conversion of ASR Transcripts
Automatic Speech Recognition (ASR) transcripts exhibit recognition errors and various spoken language phenomena such as disfluencies, ungrammatical sentences, and incomplete sentences, hence suffering from poor readability. To improve readability, we propose a Contextualized Spoken-to-Written conversion (CoS2W) task to address ASR and grammar errors and also transfer the informal text into the formal style with content preserved, utilizing contexts and auxiliary information. This task naturally matches the in-context learning capabilities of Large Language Models (LLMs). To facilitate comprehensive comparisons of various LLMs, we construct a document-level Spoken-to-Written conversion of ASR Transcripts Benchmark (SWAB) dataset. Using SWAB, we study the impact of different granularity levels on the CoS2W performance, and propose methods to exploit contexts and auxiliary information to enhance the outputs. Experimental results reveal that LLMs have the potential to excel in the CoS2W task, particularly in grammaticality and formality, our methods achieve effective understanding of contexts and auxiliary information by LLMs. We further investigate the effectiveness of using LLMs as evaluators and find that LLM evaluators show strong correlations with human evaluations on rankings of faithfulness and formality, which validates the reliability of LLM evaluators for the CoS2W task.
comment: 7 pages, 3 figures
♻ ☆ Embedding-Informed Adaptive Retrieval-Augmented Generation of Large Language Models
Retrieval-augmented large language models (LLMs) have been remarkably competent in various NLP tasks. However, it was observed by previous works that retrieval is not always helpful, especially when the LLM is already knowledgeable on the query to answer. Motivated by this, Adaptive Retrieval-Augmented Generation (ARAG) studies retrieving only when the knowledge asked by the query is absent in the LLM. Previous works of ARAG either require accessing the pre-training corpus or prompting with additional model inferences. Aiming to avoid such drawbacks, we propose to determine whether the model is knowledgeable on a query via inspecting the (contextualized) pre-trained token embeddings of LLMs. We hypothesize that such embeddings capture rich information on the model's intrinsic knowledge base, which enables an efficient way of judging the necessity to retrieve from an external corpus. Extensive experiments demonstrate our ARAG approach's superior performance across various benchmarks.
♻ ☆ SLEB: Streamlining LLMs through Redundancy Verification and Elimination of Transformer Blocks ICML 2024
Large language models (LLMs) have proven to be highly effective across various natural language processing tasks. However, their large number of parameters poses significant challenges for practical deployment. Pruning, a technique aimed at reducing the size and complexity of LLMs, offers a potential solution by removing redundant components from the network. Despite the promise of pruning, existing methods often struggle to achieve substantial end-to-end LLM inference speedup. In this paper, we introduce SLEB, a novel approach designed to streamline LLMs by eliminating redundant transformer blocks. We choose the transformer block as the fundamental unit for pruning, because LLMs exhibit block-level redundancy with high similarity between the outputs of neighboring blocks. This choice allows us to effectively enhance the processing speed of LLMs. Our experimental results demonstrate that SLEB outperforms previous LLM pruning methods in accelerating LLM inference while also maintaining superior perplexity and accuracy, making SLEB as a promising technique for enhancing the efficiency of LLMs. The code is available at: https://github.com/jiwonsong-dev/SLEB.
comment: ICML 2024
♻ ☆ Improvement in Sign Language Translation Using Text CTC Alignment
Current sign language translation (SLT) approaches often rely on gloss-based supervision with Connectionist Temporal Classification (CTC), limiting their ability to handle non-monotonic alignments between sign language video and spoken text. In this work, we propose a novel method combining joint CTC/Attention and transfer learning. The joint CTC/Attention introduces hierarchical encoding and integrates CTC with the attention mechanism during decoding, effectively managing both monotonic and non-monotonic alignments. Meanwhile, transfer learning helps bridge the modality gap between vision and language in SLT. Experimental results on two widely adopted benchmarks, RWTH-PHOENIX-Weather 2014 T and CSL-Daily, show that our method achieves results comparable to state-of-the-art and outperforms the pure-attention baseline. Additionally, this work opens a new door for future research into gloss-free SLT using text-based CTC alignment.
♻ ☆ LongVILA: Scaling Long-Context Visual Language Models for Long Videos
Long-context capability is critical for multi-modal foundation models, especially for long video understanding. We introduce LongVILA, a full-stack solution for long-context visual-language models by co-designing the algorithm and system. For model training, we upgrade existing VLMs to support long video understanding by incorporating two additional stages, i.e., long context extension and long video supervised fine-tuning. However, training on long video is computationally and memory intensive. We introduce the long-context Multi-Modal Sequence Parallelism (MM-SP) system that efficiently parallelizes long video training and inference, enabling 2M context length training on 256 GPUs without any gradient checkpointing. LongVILA efficiently extends the number of video frames of VILA from 8 to 2048, achieving 99.8% accuracy in 6,000-frame (more than 1 million tokens) video needle-in-a-haystack. LongVILA-7B demonstrates strong accuracy on 9 popular video benchmarks, e.g. 65.1% VideoMME with subtitle. Besides, MM-SP is 2.1x - 5.7x faster than ring style sequence parallelism and 1.1x - 1.4x faster than Megatron with a hybrid context and tensor parallelism. Moreover, it seamlessly integrates with Hugging Face Transformers.
comment: Code and models are available at https://github.com/NVlabs/VILA/tree/main/longvila
Information Retrieval 4
☆ MST-R: Multi-Stage Tuning for Retrieval Systems and Metric Evaluation
Regulatory documents are rich in nuanced terminology and specialized semantics. FRAG systems: Frozen retrieval-augmented generators utilizing pre-trained (or, frozen) components face consequent challenges with both retriever and answering performance. We present a system that adapts the retriever performance to the target domain using a multi-stage tuning (MST) strategy. Our retrieval approach, called MST-R (a) first fine-tunes encoders used in vector stores using hard negative mining, (b) then uses a hybrid retriever, combining sparse and dense retrievers using reciprocal rank fusion, and then (c) adapts the cross-attention encoder by fine-tuning only the top-k retrieved results. We benchmark the system performance on the dataset released for the RIRAG challenge (as part of the RegNLP workshop at COLING 2025). We achieve significant performance gains obtaining a top rank on the RegNLP challenge leaderboard. We also show that a trivial answering approach games the RePASs metric outscoring all baselines and a pre-trained Llama model. Analyzing this anomaly, we present important takeaways for future research.
☆ Static Pruning in Dense Retrieval using Matrix Decomposition
In the era of dense retrieval, document indexing and retrieval is largely based on encoding models that transform text documents into embeddings. The efficiency of retrieval is directly proportional to the number of documents and the size of the embeddings. Recent studies have shown that it is possible to reduce embedding size without sacrificing - and in some cases improving - the retrieval effectiveness. However, the methods introduced by these studies are query-dependent, so they can't be applied offline and require additional computations during query processing, thus negatively impacting the retrieval efficiency. In this paper, we present a novel static pruning method for reducing the dimensionality of embeddings using Principal Components Analysis. This approach is query-independent and can be executed offline, leading to a significant boost in dense retrieval efficiency with a negligible impact on the system effectiveness. Our experiments show that our proposed method reduces the dimensionality of document representations by over 50% with up to a 5% reduction in NDCG@10, for different dense retrieval models.
☆ Hesitation and Tolerance in Recommender Systems
User interactions in recommender systems are inherently complex, often involving behaviors that go beyond simple acceptance or rejection. One particularly common behavior is hesitation, where users deliberate over recommended items, signaling uncertainty. Our large-scale surveys, with 6,644 and 3,864 responses respectively, confirm that hesitation is not only widespread but also has a profound impact on user experiences. When users spend additional time engaging with content they are ultimately uninterested in, this can lead to negative emotions, a phenomenon we term as tolerance. The surveys reveal that such tolerance behaviors often arise after hesitation and can erode trust, satisfaction, and long-term loyalty to the platform. For instance, a click might reflect a need for more information rather than genuine interest, and prolonged exposure to unsuitable content amplifies frustration. This misalignment between user intent and system interpretation introduces noise into recommendation training, resulting in suggestions that increase uncertainty and disengagement. To address these issues, we identified signals indicative of tolerance behavior and analyzed datasets from both e-commerce and short-video platforms. The analysis shows a strong correlation between increased tolerance behavior and decreased user activity. We integrated these insights into the training process of a recommender system for a major short-video platform. Results from four independent online A/B experiments demonstrated significant improvements in user retention, achieved with minimal additional computational costs. These findings underscore the importance of recognizing hesitation as a ubiquitous user behavior and addressing tolerance to enhance satisfaction, build trust, and sustain long-term engagement in recommender systems.
comment: 30 pages, 6 figures, 6 tables
♻ ☆ Navigating the Future of Federated Recommendation Systems with Foundation Models
In recent years, the integration of federated learning (FL) and recommendation systems (RS), known as Federated Recommendation Systems (FRS), has attracted attention for preserving user privacy by keeping private data on client devices. However, FRS faces inherent limitations such as data heterogeneity and scarcity, due to the privacy requirements of FL and the typical data sparsity issues of RSs. Models like ChatGPT are empowered by the concept of transfer learning and self-supervised learning, so they can be easily applied to the downstream tasks after fine-tuning or prompting. These models, so-called Foundation Models (FM), fouce on understanding the human's intent and perform following their designed roles in the specific tasks, which are widely recognized for producing high-quality content in the image and language domains. Thus, the achievements of FMs inspire the design of FRS and suggest a promising research direction: integrating foundation models to address the above limitations. In this study, we conduct a comprehensive review of FRSs with FMs. Specifically, we: 1) summarise the common approaches of current FRSs and FMs; 2) review the challenges posed by FRSs and FMs; 3) discuss potential future research directions; and 4) introduce some common benchmarks and evaluation metrics in the FRS field. We hope that this position paper provides the necessary background and guidance to explore this interesting and emerging topic.
comment: 20 pages, position paper, survey
Machine Learning 139
☆ GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Prediction
3D occupancy prediction is important for autonomous driving due to its comprehensive perception of the surroundings. To incorporate sequential inputs, most existing methods fuse representations from previous frames to infer the current 3D occupancy. However, they fail to consider the continuity of driving scenarios and ignore the strong prior provided by the evolution of 3D scenes (e.g., only dynamic objects move). In this paper, we propose a world-model-based framework to exploit the scene evolution for perception. We reformulate 3D occupancy prediction as a 4D occupancy forecasting problem conditioned on the current sensor input. We decompose the scene evolution into three factors: 1) ego motion alignment of static scenes; 2) local movements of dynamic objects; and 3) completion of newly-observed scenes. We then employ a Gaussian world model (GaussianWorld) to explicitly exploit these priors and infer the scene evolution in the 3D Gaussian space considering the current RGB observation. We evaluate the effectiveness of our framework on the widely used nuScenes dataset. Our GaussianWorld improves the performance of the single-frame counterpart by over 2% in mIoU without introducing additional computations. Code: https://github.com/zuosc19/GaussianWorld.
comment: Code is available at: https://github.com/zuosc19/GaussianWorld
☆ GaussianAD: Gaussian-Centric End-to-End Autonomous Driving
Vision-based autonomous driving shows great potential due to its satisfactory performance and low costs. Most existing methods adopt dense representations (e.g., bird's eye view) or sparse representations (e.g., instance boxes) for decision-making, which suffer from the trade-off between comprehensiveness and efficiency. This paper explores a Gaussian-centric end-to-end autonomous driving (GaussianAD) framework and exploits 3D semantic Gaussians to extensively yet sparsely describe the scene. We initialize the scene with uniform 3D Gaussians and use surrounding-view images to progressively refine them to obtain the 3D Gaussian scene representation. We then use sparse convolutions to efficiently perform 3D perception (e.g., 3D detection, semantic map construction). We predict 3D flows for the Gaussians with dynamic semantics and plan the ego trajectory accordingly with an objective of future scene forecasting. Our GaussianAD can be trained in an end-to-end manner with optional perception labels when available. Extensive experiments on the widely used nuScenes dataset verify the effectiveness of our end-to-end GaussianAD on various tasks including motion planning, 3D occupancy prediction, and 4D occupancy forecasting. Code: https://github.com/wzzheng/GaussianAD.
comment: Code is available at: https://github.com/wzzheng/GaussianAD
☆ OP-LoRA: The Blessing of Dimensionality
Low-rank adapters enable fine-tuning of large models with only a small number of parameters, thus reducing storage costs and minimizing the risk of catastrophic forgetting. However, they often pose optimization challenges, with poor convergence. To overcome these challenges, we introduce an over-parameterized approach that accelerates training without increasing inference costs. This method reparameterizes low-rank adaptation by employing a separate MLP and learned embedding for each layer. The learned embedding is input to the MLP, which generates the adapter parameters. Such overparamaterization has been shown to implicitly function as an adaptive learning rate and momentum, accelerating optimization. At inference time, the MLP can be discarded, leaving behind a standard low-rank adapter. To study the effect of MLP overparameterization on a small yet difficult proxy task, we implement it for matrix factorization, and find it achieves faster convergence and lower final loss. Extending this approach to larger-scale tasks, we observe consistent performance gains across domains. We achieve improvements in vision-language tasks and especially notable increases in image generation, with CMMD scores improving by up to 15 points.
☆ The Correlated Gaussian Sparse Histogram Mechanism
We consider the problem of releasing a sparse histogram under $(\varepsilon, \delta)$-differential privacy. The stability histogram independently adds noise from a Laplace or Gaussian distribution to the non-zero entries and removes those noisy counts below a threshold. Thereby, the introduction of new non-zero values between neighboring histograms is only revealed with probability at most $\delta$, and typically, the value of the threshold dominates the error of the mechanism. We consider the variant of the stability histogram with Gaussian noise. Recent works ([Joseph and Yu, COLT '24] and [Lebeda, SOSA '25]) reduced the error for private histograms using correlated Gaussian noise. However, these techniques can not be directly applied in the very sparse setting. Instead, we adopt Lebeda's technique and show that adding correlated noise to the non-zero counts only allows us to reduce the magnitude of noise when we have a sparsity bound. This, in turn, allows us to use a lower threshold by up to a factor of $1/2$ compared to the non-correlated noise mechanism. We then extend our mechanism to a setting without a known bound on sparsity. Additionally, we show that correlated noise can give a similar improvement for the more practical discrete Gaussian mechanism.
☆ A Library for Learning Neural Operators
We present NeuralOperator, an open-source Python library for operator learning. Neural operators generalize neural networks to maps between function spaces instead of finite-dimensional Euclidean spaces. They can be trained and inferenced on input and output functions given at various discretizations, satisfying a discretization convergence properties. Built on top of PyTorch, NeuralOperator provides all the tools for training and deploying neural operator models, as well as developing new ones, in a high-quality, tested, open-source package. It combines cutting-edge models and customizability with a gentle learning curve and simple user interface for newcomers.
☆ Robust image classification with multi-modal large language models
Deep Neural Networks are vulnerable to adversarial examples, i.e., carefully crafted input samples that can cause models to make incorrect predictions with high confidence. To mitigate these vulnerabilities, adversarial training and detection-based defenses have been proposed to strengthen models in advance. However, most of these approaches focus on a single data modality, overlooking the relationships between visual patterns and textual descriptions of the input. In this paper, we propose a novel defense, Multi-Shield, designed to combine and complement these defenses with multi-modal information to further enhance their robustness. Multi-Shield leverages multi-modal large language models to detect adversarial examples and abstain from uncertain classifications when there is no alignment between textual and visual representations of the input. Extensive evaluations on CIFAR-10 and ImageNet datasets, using robust and non-robust image classification models, demonstrate that Multi-Shield can be easily integrated to detect and reject adversarial examples, outperforming the original defenses.
☆ COMET: Benchmark for Comprehensive Biological Multi-omics Evaluation Tasks and Language Models
As key elements within the central dogma, DNA, RNA, and proteins play crucial roles in maintaining life by guaranteeing accurate genetic expression and implementation. Although research on these molecules has profoundly impacted fields like medicine, agriculture, and industry, the diversity of machine learning approaches-from traditional statistical methods to deep learning models and large language models-poses challenges for researchers in choosing the most suitable models for specific tasks, especially for cross-omics and multi-omics tasks due to the lack of comprehensive benchmarks. To address this, we introduce the first comprehensive multi-omics benchmark COMET (Benchmark for Biological COmprehensive Multi-omics Evaluation Tasks and Language Models), designed to evaluate models across single-omics, cross-omics, and multi-omics tasks. First, we curate and develop a diverse collection of downstream tasks and datasets covering key structural and functional aspects in DNA, RNA, and proteins, including tasks that span multiple omics levels. Then, we evaluate existing foundational language models for DNA, RNA, and proteins, as well as the newly proposed multi-omics method, offering valuable insights into their performance in integrating and analyzing data from different biological modalities. This benchmark aims to define critical issues in multi-omics research and guide future directions, ultimately promoting advancements in understanding biological processes through integrated and different omics data analysis.
☆ Shape error prediction in 5-axis machining using graph neural networks
This paper presents an innovative method for predicting shape errors in 5-axis machining using graph neural networks. The graph structure is defined with nodes representing workpiece surface points and edges denoting the neighboring relationships. The dataset encompasses data from a material removal simulation, process data, and post-machining quality information. Experimental results show that the presented approach can generalize the shape error prediction for the investigated workpiece geometry. Moreover, by modelling spatial and temporal connections within the workpiece, the approach handles a low number of labels compared to non-graphical methods such as Support Vector Machines.
☆ Generative AI in Medicine
The increased capabilities of generative AI have dramatically expanded its possible use cases in medicine. We provide a comprehensive overview of generative AI use cases for clinicians, patients, clinical trial organizers, researchers, and trainees. We then discuss the many challenges -- including maintaining privacy and security, improving transparency and interpretability, upholding equity, and rigorously evaluating models -- which must be overcome to realize this potential, and the open research directions they give rise to.
comment: To appear in the Annual Review of Biomedical Data Science, August 2025
☆ AdvPrefix: An Objective for Nuanced LLM Jailbreaks
Many jailbreak attacks on large language models (LLMs) rely on a common objective: making the model respond with the prefix "Sure, here is (harmful request)". While straightforward, this objective has two limitations: limited control over model behaviors, often resulting in incomplete or unrealistic responses, and a rigid format that hinders optimization. To address these limitations, we introduce AdvPrefix, a new prefix-forcing objective that enables more nuanced control over model behavior while being easy to optimize. Our objective leverages model-dependent prefixes, automatically selected based on two criteria: high prefilling attack success rates and low negative log-likelihood. It can further simplify optimization by using multiple prefixes for a single user request. AdvPrefix can integrate seamlessly into existing jailbreak attacks to improve their performance for free. For example, simply replacing GCG attack's target prefixes with ours on Llama-3 improves nuanced attack success rates from 14% to 80%, suggesting that current alignment struggles to generalize to unseen prefixes. Our work demonstrates the importance of jailbreak objectives in achieving nuanced jailbreaks.
☆ SCBench: A KV Cache-Centric Analysis of Long-Context Methods
Long-context LLMs have enabled numerous downstream applications but also introduced significant challenges related to computational and memory efficiency. To address these challenges, optimizations for long-context inference have been developed, centered around the KV cache. However, existing benchmarks often evaluate in single-request, neglecting the full lifecycle of the KV cache in real-world use. This oversight is particularly critical, as KV cache reuse has become widely adopted in LLMs inference frameworks, such as vLLM and SGLang, as well as by LLM providers, including OpenAI, Microsoft, Google, and Anthropic. To address this gap, we introduce SCBench(SharedContextBench), a comprehensive benchmark for evaluating long-context methods from a KV cachecentric perspective: 1) KV cache generation, 2) KV cache compression, 3) KV cache retrieval, 4) KV cache loading. Specifically, SCBench uses test examples with shared context, ranging 12 tasks with two shared context modes, covering four categories of long-context capabilities: string retrieval, semantic retrieval, global information, and multi-task. With it, we provide an extensive KV cache-centric analysis of eight categories long-context solutions, including Gated Linear RNNs, Mamba-Attention hybrids, and efficient methods such as sparse attention, KV cache dropping, quantization, retrieval, loading, and prompt compression. The evaluation is conducted on 8 long-context LLMs. Our findings show that sub-O(n) memory methods suffer in multi-turn scenarios, while sparse encoding with O(n) memory and sub-O(n^2) pre-filling computation perform robustly. Dynamic sparsity yields more expressive KV caches than static patterns, and layer-level sparsity in hybrid architectures reduces memory usage with strong performance. Additionally, we identify attention distribution shift issues in long-generation scenarios. https://aka.ms/SCBench.
☆ MST-R: Multi-Stage Tuning for Retrieval Systems and Metric Evaluation
Regulatory documents are rich in nuanced terminology and specialized semantics. FRAG systems: Frozen retrieval-augmented generators utilizing pre-trained (or, frozen) components face consequent challenges with both retriever and answering performance. We present a system that adapts the retriever performance to the target domain using a multi-stage tuning (MST) strategy. Our retrieval approach, called MST-R (a) first fine-tunes encoders used in vector stores using hard negative mining, (b) then uses a hybrid retriever, combining sparse and dense retrievers using reciprocal rank fusion, and then (c) adapts the cross-attention encoder by fine-tuning only the top-k retrieved results. We benchmark the system performance on the dataset released for the RIRAG challenge (as part of the RegNLP workshop at COLING 2025). We achieve significant performance gains obtaining a top rank on the RegNLP challenge leaderboard. We also show that a trivial answering approach games the RePASs metric outscoring all baselines and a pre-trained Llama model. Analyzing this anomaly, we present important takeaways for future research.
☆ Interlocking-free Selective Rationalization Through Genetic-based Learning
A popular end-to-end architecture for selective rationalization is the select-then-predict pipeline, comprising a generator to extract highlights fed to a predictor. Such a cooperative system suffers from suboptimal equilibrium minima due to the dominance of one of the two modules, a phenomenon known as interlocking. While several contributions aimed at addressing interlocking, they only mitigate its effect, often by introducing feature-based heuristics, sampling, and ad-hoc regularizations. We present GenSPP, the first interlocking-free architecture for selective rationalization that does not require any learning overhead, as the above-mentioned. GenSPP avoids interlocking by performing disjoint training of the generator and predictor via genetic global search. Experiments on a synthetic and a real-world benchmark show that our model outperforms several state-of-the-art competitors.
☆ Buzz to Broadcast: Predicting Sports Viewership Using Social Media Engagement
Accurately predicting sports viewership is crucial for optimizing ad sales and revenue forecasting. Social media platforms, such as Reddit, provide a wealth of user-generated content that reflects audience engagement and interest. In this study, we propose a regression-based approach to predict sports viewership using social media metrics, including post counts, comments, scores, and sentiment analysis from TextBlob and VADER. Through iterative improvements, such as focusing on major sports subreddits, incorporating categorical features, and handling outliers by sport, the model achieved an $R^2$ of 0.99, a Mean Absolute Error (MAE) of 1.27 million viewers, and a Root Mean Squared Error (RMSE) of 2.33 million viewers on the full dataset. These results demonstrate the model's ability to accurately capture patterns in audience behavior, offering significant potential for pre-event revenue forecasting and targeted advertising strategies.
comment: 17 pages, 4 figures, 2 tables
☆ Still "Talking About Large Language Models": Some Clarifications
My paper "Talking About Large Language Models" has more than once been interpreted as advocating a reductionist stance towards large language models. But the paper was not intended that way, and I do not endorse such positions. This short note situates the paper in the context of a larger philosophical project that is concerned with the (mis)use of words rather than metaphysics, in the spirit of Wittgenstein's later writing.
☆ Performance evaluation of predictive AI models to support medical decisions: Overview and guidance
A myriad of measures to illustrate performance of predictive artificial intelligence (AI) models have been proposed in the literature. Selecting appropriate performance measures is essential for predictive AI models that are developed to be used in medical practice, because poorly performing models may harm patients and lead to increased costs. We aim to assess the merits of classic and contemporary performance measures when validating predictive AI models for use in medical practice. We focus on models with a binary outcome. We discuss 32 performance measures covering five performance domains (discrimination, calibration, overall, classification, and clinical utility) along with accompanying graphical assessments. The first four domains cover statistical performance, the fifth domain covers decision-analytic performance. We explain why two key characteristics are important when selecting which performance measures to assess: (1) whether the measure's expected value is optimized when it is calculated using the correct probabilities (i.e., a "proper" measure), and (2) whether they reflect either purely statistical performance or decision-analytic performance by properly considering misclassification costs. Seventeen measures exhibit both characteristics, fourteen measures exhibited one characteristic, and one measure possessed neither characteristic (the F1 measure). All classification measures (such as classification accuracy and F1) are improper for clinically relevant decision thresholds other than 0.5 or the prevalence. We recommend the following measures and plots as essential to report: AUROC, calibration plot, a clinical utility measure such as net benefit with decision curve analysis, and a plot with probability distributions per outcome category.
comment: 60 pages, 8 tables, 11 figures, two supplementary appendices
☆ Probabilistic Inverse Cameras: Image to 3D via Multiview Geometry
We introduce a hierarchical probabilistic approach to go from a 2D image to multiview 3D: a diffusion "prior" models the unseen 3D geometry, which then conditions a diffusion "decoder" to generate novel views of the subject. We use a pointmap-based geometric representation in a multiview image format to coordinate the generation of multiple target views simultaneously. We facilitate correspondence between views by assuming fixed target camera poses relative to the source camera, and constructing a predictable distribution of geometric features per target. Our modular, geometry-driven approach to novel-view synthesis (called "unPIC") beats SoTA baselines such as CAT3D and One-2-3-45 on held-out objects from ObjaverseXL, as well as real-world objects ranging from Google Scanned Objects, Amazon Berkeley Objects, to the Digital Twin Catalog.
☆ Adversarial Robustness of Bottleneck Injected Deep Neural Networks for Task-Oriented Communication ICML
This paper investigates the adversarial robustness of Deep Neural Networks (DNNs) using Information Bottleneck (IB) objectives for task-oriented communication systems. We empirically demonstrate that while IB-based approaches provide baseline resilience against attacks targeting downstream tasks, the reliance on generative models for task-oriented communication introduces new vulnerabilities. Through extensive experiments on several datasets, we analyze how bottleneck depth and task complexity influence adversarial robustness. Our key findings show that Shallow Variational Bottleneck Injection (SVBI) provides less adversarial robustness compared to Deep Variational Information Bottleneck (DVIB) approaches, with the gap widening for more complex tasks. Additionally, we reveal that IB-based objectives exhibit stronger robustness against attacks focusing on salient pixels with high intensity compared to those perturbing many pixels with lower intensity. Lastly, we demonstrate that task-oriented communication systems that rely on generative models to extract and recover salient information have an increased attack surface. The results highlight important security considerations for next-generation communication systems that leverage neural networks for goal-oriented compression.
comment: Submission to ICMLCN, 6 pages, 9 figures, 3 tables
☆ Controlling dynamical systems into unseen target states using machine learning
We present a novel, model-free, and data-driven methodology for controlling complex dynamical systems into previously unseen target states, including those with significantly different and complex dynamics. Leveraging a parameter-aware realization of next-generation reservoir computing, our approach accurately predicts system behavior in unobserved parameter regimes, enabling control over transitions to arbitrary target states. Crucially, this includes states with dynamics that differ fundamentally from known regimes, such as shifts from periodic to intermittent or chaotic behavior. The method's parameter-awareness facilitates non-stationary control, ensuring smooth transitions between states. By extending the applicability of machine learning-based control mechanisms to previously inaccessible target dynamics, this methodology opens the door to transformative new applications while maintaining exceptional efficiency. Our results highlight reservoir computing as a powerful alternative to traditional methods for dynamic system control.
Detecting LLM Hallucination Through Layer-wise Information Deficiency: Analysis of Unanswerable Questions and Ambiguous Prompts
Large language models (LLMs) frequently generate confident yet inaccurate responses, introducing significant risks for deployment in safety-critical domains. We present a novel approach to detecting model hallucination through systematic analysis of information flow across model layers when processing inputs with insufficient or ambiguous context. Our investigation reveals that hallucination manifests as usable information deficiencies in inter-layer transmissions. While existing approaches primarily focus on final-layer output analysis, we demonstrate that tracking cross-layer information dynamics ($\mathcal{L}$I) provides robust indicators of model reliability, accounting for both information gain and loss during computation. $\mathcal{L}$I improves model reliability by immediately integrating with universal LLMs without additional training or architectural modifications.
☆ Efficient Continual Pre-training of LLMs for Low-resource Languages
Open-source Large Language models (OsLLMs) propel the democratization of natural language research by giving the flexibility to augment or update model parameters for performance improvement. Nevertheless, like proprietary LLMs, Os-LLMs offer poorer performance on low-resource languages (LRLs) than high-resource languages (HRLs), owing to smaller amounts of training data and underrepresented vocabulary. On the other hand, continual pre-training (CPT) with large amounts of language-specific data is a costly proposition in terms of data acquisition and computational resources. Our goal is to drastically reduce CPT cost. To that end, we first develop a new algorithm to select a subset of texts from a larger corpus. We show the effectiveness of our technique using very little CPT data. In search of further improvement, we design a new algorithm to select tokens to include in the LLM vocabulary. We experiment with the recent Llama-3 model and nine Indian languages with diverse scripts and extent of resource availability. For evaluation, we use IndicGenBench, a generation task benchmark dataset for Indic languages. We experiment with various CPT corpora and augmented vocabulary size and offer insights across language families.
☆ Efficient Generative Modeling with Residual Vector Quantization-Based Tokens
We explore the use of Residual Vector Quantization (RVQ) for high-fidelity generation in vector-quantized generative models. This quantization technique maintains higher data fidelity by employing more in-depth tokens. However, increasing the token number in generative models leads to slower inference speeds. To this end, we introduce ResGen, an efficient RVQ-based discrete diffusion model that generates high-fidelity samples without compromising sampling speed. Our key idea is a direct prediction of vector embedding of collective tokens rather than individual ones. Moreover, we demonstrate that our proposed token masking and multi-token prediction method can be formulated within a principled probabilistic framework using a discrete diffusion process and variational inference. We validate the efficacy and generalizability of the proposed method on two challenging tasks across different modalities: conditional image generation} on ImageNet 256x256 and zero-shot text-to-speech synthesis. Experimental results demonstrate that ResGen outperforms autoregressive counterparts in both tasks, delivering superior performance without compromising sampling speed. Furthermore, as we scale the depth of RVQ, our generative models exhibit enhanced generation fidelity or faster sampling speeds compared to similarly sized baseline models. The project page can be found at https://resgen-genai.github.io
☆ Integrative Analysis of Financial Market Sentiment Using CNN and GRU for Risk Prediction and Alert Systems
This document presents an in-depth examination of stock market sentiment through the integration of Convolutional Neural Networks (CNN) and Gated Recurrent Units (GRU), enabling precise risk alerts. The robust feature extraction capability of CNN is utilized to preprocess and analyze extensive network text data, identifying local features and patterns. The extracted feature sequences are then input into the GRU model to understand the progression of emotional states over time and their potential impact on future market sentiment and risk. This approach addresses the order dependence and long-term dependencies inherent in time series data, resulting in a detailed analysis of stock market sentiment and effective early warnings of future risks.
☆ Simple Guidance Mechanisms for Discrete Diffusion Models
Diffusion models for continuous data gained widespread adoption owing to their high quality generation and control mechanisms. However, controllable diffusion on discrete data faces challenges given that continuous guidance methods do not directly apply to discrete diffusion. Here, we provide a straightforward derivation of classifier-free and classifier-based guidance for discrete diffusion, as well as a new class of diffusion models that leverage uniform noise and that are more guidable because they can continuously edit their outputs. We improve the quality of these models with a novel continuous-time variational lower bound that yields state-of-the-art performance, especially in settings involving guidance or fast generation. Empirically, we demonstrate that our guidance mechanisms combined with uniform noise diffusion improve controllable generation relative to autoregressive and diffusion baselines on several discrete data domains, including genomic sequences, small molecule design, and discretized image generation.
comment: Code to reproduce our experiments is available here: https://github.com/kuleshov-group/discrete-diffusion-guidance
☆ BiCert: A Bilinear Mixed Integer Programming Formulation for Precise Certified Bounds Against Data Poisoning Attacks
Data poisoning attacks pose one of the biggest threats to modern AI systems, necessitating robust defenses. While extensive efforts have been made to develop empirical defenses, attackers continue to evolve, creating sophisticated methods to circumvent these measures. To address this, we must move beyond empirical defenses and establish provable certification methods that guarantee robustness. This paper introduces a novel certification approach, BiCert, using Bilinear Mixed Integer Programming (BMIP) to compute sound deterministic bounds that provide such provable robustness. Using BMIP, we compute the reachable set of parameters that could result from training with potentially manipulated data. A key element to make this computation feasible is to relax the reachable parameter set to a convex set between training iterations. At test time, this parameter set allows us to predict all possible outcomes, guaranteeing robustness. BiCert is more precise than previous methods, which rely solely on interval and polyhedral bounds. Crucially, our approach overcomes the fundamental limitation of prior approaches where parameter bounds could only grow, often uncontrollably. We show that BiCert's tighter bounds eliminate a key source of divergence issues, resulting in more stable training and higher certified accuracy.
☆ Solving Robust Markov Decision Processes: Generic, Reliable, Efficient AAAI'25
Markov decision processes (MDP) are a well-established model for sequential decision-making in the presence of probabilities. In robust MDP (RMDP), every action is associated with an uncertainty set of probability distributions, modelling that transition probabilities are not known precisely. Based on the known theoretical connection to stochastic games, we provide a framework for solving RMDPs that is generic, reliable, and efficient. It is *generic* both with respect to the model, allowing for a wide range of uncertainty sets, including but not limited to intervals, $L^1$- or $L^2$-balls, and polytopes; and with respect to the objective, including long-run average reward, undiscounted total reward, and stochastic shortest path. It is *reliable*, as our approach not only converges in the limit, but provides precision guarantees at any time during the computation. It is *efficient* because -- in contrast to state-of-the-art approaches -- it avoids explicitly constructing the underlying stochastic game. Consequently, our prototype implementation outperforms existing tools by several orders of magnitude and can solve RMDPs with a million states in under a minute.
comment: Accepted for publication at AAAI'25. Extended version with full appendix, 26 pages
☆ Sims: An Interactive Tool for Geospatial Matching and Clustering
Acquiring, processing, and visualizing geospatial data requires significant computing resources, especially for large spatio-temporal domains. This challenge hinders the rapid discovery of predictive features, which is essential for advancing geospatial modeling. To address this, we developed Similarity Search (Sims), a no-code web tool that allows users to visualize, compare, cluster, and perform similarity search over defined regions of interest using Google Earth Engine as a backend. Sims is designed to complement existing modeling tools by focusing on feature exploration rather than model creation. We demonstrate the utility of Sims through a case study analyzing simulated maize yield data in Rwanda, where we evaluate how different combinations of soil, weather, and agronomic features affect the clustering of yield response zones. Sims is open source and available at https://github.com/microsoft/Sims
☆ Multi-Head Encoding for Extreme Label Classification
The number of categories of instances in the real world is normally huge, and each instance may contain multiple labels. To distinguish these massive labels utilizing machine learning, eXtreme Label Classification (XLC) has been established. However, as the number of categories increases, the number of parameters and nonlinear operations in the classifier also rises. This results in a Classifier Computational Overload Problem (CCOP). To address this, we propose a Multi-Head Encoding (MHE) mechanism, which replaces the vanilla classifier with a multi-head classifier. During the training process, MHE decomposes extreme labels into the product of multiple short local labels, with each head trained on these local labels. During testing, the predicted labels can be directly calculated from the local predictions of each head. This reduces the computational load geometrically. Then, according to the characteristics of different XLC tasks, e.g., single-label, multi-label, and model pretraining tasks, three MHE-based implementations, i.e., Multi-Head Product, Multi-Head Cascade, and Multi-Head Sampling, are proposed to more effectively cope with CCOP. Moreover, we theoretically demonstrate that MHE can achieve performance approximately equivalent to that of the vanilla classifier by generalizing the low-rank approximation problem from Frobenius-norm to Cross-Entropy. Experimental results show that the proposed methods achieve state-of-the-art performance while significantly streamlining the training and inference processes of XLC tasks. The source code has been made public at https://github.com/Anoise/MHE.
comment: 20 pages, 12 figs, Published in TPAMI
☆ Learning payoffs while routing in skill-based queues
Motivated by applications in service systems, we consider queueing systems where each customer must be handled by a server with the right skill set. We focus on optimizing the routing of customers to servers in order to maximize the total payoff of customer--server matches. In addition, customer--server dependent payoff parameters are assumed to be unknown a priori. We construct a machine learning algorithm that adaptively learns the payoff parameters while maximizing the total payoff and prove that it achieves polylogarithmic regret. Moreover, we show that the algorithm is asymptotically optimal up to logarithmic terms by deriving a regret lower bound. The algorithm leverages the basic feasible solutions of a static linear program as the action space. The regret analysis overcomes the complex interplay between queueing and learning by analyzing the convergence of the queue length process to its stationary behavior. We also demonstrate the performance of the algorithm numerically, and have included an experiment with time-varying parameters highlighting the potential of the algorithm in non-static environments.
☆ Scaling Combinatorial Optimization Neural Improvement Heuristics with Online Search and Adaptation
We introduce Limited Rollout Beam Search (LRBS), a beam search strategy for deep reinforcement learning (DRL) based combinatorial optimization improvement heuristics. Utilizing pre-trained models on the Euclidean Traveling Salesperson Problem, LRBS significantly enhances both in-distribution performance and generalization to larger problem instances, achieving optimality gaps that outperform existing improvement heuristics and narrowing the gap with state-of-the-art constructive methods. We also extend our analysis to two pickup and delivery TSP variants to validate our results. Finally, we employ our search strategy for offline and online adaptation of the pre-trained improvement policy, leading to improved search performance and surpassing recent adaptive methods for constructive heuristics.
☆ Data Integration with Fusion Searchlight: Classifying Brain States from Resting-state fMRI
Spontaneous neural activity observed in resting-state fMRI is characterized by complex spatio-temporal dynamics. Different measures related to local and global brain connectivity and fluctuations in low-frequency amplitudes can quantify individual aspects of these neural dynamics. Even though such measures are derived from the same functional signals, they are often evaluated separately, neglecting their interrelations and potentially reducing the analysis sensitivity. In our study, we present a fusion searchlight (FuSL) framework to combine the complementary information contained in different resting-state fMRI metrics and demonstrate how this can improve the decoding of brain states. Moreover, we show how explainable AI allows us to reconstruct the differential impact of each metric on the decoding, which additionally increases spatial specificity of searchlight analysis. In general, this framework can be adapted to combine information derived from different imaging modalities or experimental conditions, offering a versatile and interpretable tool for data fusion in neuroimaging.
☆ Investigating generalization capabilities of neural networks by means of loss landscapes and Hessian analysis
This paper studies generalization capabilities of neural networks (NNs) using new and improved PyTorch library Loss Landscape Analysis (LLA). LLA facilitates visualization and analysis of loss landscapes along with the properties of NN Hessian. Different approaches to NN loss landscape plotting are discussed with particular focus on normalization techniques showing that conventional methods cannot always ensure correct visualization when batch normalization layers are present in NN architecture. The use of Hessian axes is shown to be able to mitigate this effect, and methods for choosing Hessian axes are proposed. In addition, spectra of Hessian eigendecomposition are studied and it is shown that typical spectra exist for a wide range of NNs. This allows to propose quantitative criteria for Hessian analysis that can be applied to evaluate NN performance and assess its generalization capabilities. Generalization experiments are conducted using ImageNet-1K pre-trained models along with several models trained as part of this study. The experiment include training models on one dataset and testing on another one to maximize experiment similarity to model performance in the Wild. It is shown that when datasets change, the changes in criteria correlate with the changes in accuracy, making the proposed criteria a computationally efficient estimate of generalization ability, which is especially useful for extremely large datasets.
☆ Can LLMs Convert Graphs to Text-Attributed Graphs?
Graphs are ubiquitous data structures found in numerous real-world applications, such as drug discovery, recommender systems, and social network analysis. Graph neural networks (GNNs) have become a popular tool to learn node embeddings through message passing on these structures. However, a significant challenge arises when applying GNNs to multiple graphs with different feature spaces, as existing GNN architectures are not designed for cross-graph feature alignment. To address this, recent approaches introduce text-attributed graphs, where each node is associated with a textual description, enabling the use of a shared textual encoder to project nodes from different graphs into a unified feature space. While promising, this method relies heavily on the availability of text-attributed data, which can be difficult to obtain in practice. To bridge this gap, we propose a novel method named Topology-Aware Node description Synthesis (TANS), which leverages large language models (LLMs) to automatically convert existing graphs into text-attributed graphs. The key idea is to integrate topological information with each node's properties, enhancing the LLMs' ability to explain how graph topology influences node semantics. We evaluate our TANS on text-rich, text-limited, and text-free graphs, demonstrating that it enables a single GNN to operate across diverse graphs. Notably, on text-free graphs, our method significantly outperforms existing approaches that manually design node features, showcasing the potential of LLMs for preprocessing graph-structured data, even in the absence of textual information. The code and data are available at https://github.com/Zehong-Wang/TANS.
☆ Optimal Bounds for Private Minimum Spanning Trees via Input Perturbation
We study the problem of privately releasing an approximate minimum spanning tree (MST). Given a graph $G = (V, E, \vec{W})$ where $V$ is a set of $n$ vertices, $E$ is a set of $m$ undirected edges, and $ \vec{W} \in \mathbb{R}^{|E|} $ is an edge-weight vector, our goal is to publish an approximate MST under edge-weight differential privacy, as introduced by Sealfon in PODS 2016, where $V$ and $E$ are considered public and the weight vector is private. Our neighboring relation is $\ell_\infty$-distance on weights: for a sensitivity parameter $\Delta_\infty$, graphs $ G = (V, E, \vec{W}) $ and $ G' = (V, E, \vec{W}') $ are neighboring if $\|\vec{W}-\vec{W}'\|_\infty \leq \Delta_\infty$. Existing private MST algorithms face a trade-off, sacrificing either computational efficiency or accuracy. We show that it is possible to get the best of both worlds: With a suitable random perturbation of the input that does not suffice to make the weight vector private, the result of any non-private MST algorithm will be private and achieves a state-of-the-art error guarantee. Furthermore, by establishing a connection to Private Top-k Selection [Steinke and Ullman, FOCS '17], we give the first privacy-utility trade-off lower bound for MST under approximate differential privacy, demonstrating that the error magnitude, $\tilde{O}(n^{3/2})$, is optimal up to logarithmic factors. That is, our approach matches the time complexity of any non-private MST algorithm and at the same time achieves optimal error. We complement our theoretical treatment with experiments that confirm the practicality of our approach.
☆ Feature Selection for Latent Factor Models CVPR
Feature selection is crucial for pinpointing relevant features in high-dimensional datasets, mitigating the 'curse of dimensionality,' and enhancing machine learning performance. Traditional feature selection methods for classification use data from all classes to select features for each class. This paper explores feature selection methods that select features for each class separately, using class models based on low-rank generative methods and introducing a signal-to-noise ratio (SNR) feature selection criterion. This novel approach has theoretical true feature recovery guarantees under certain assumptions and is shown to outperform some existing feature selection methods on standard classification datasets.
comment: Submitted to the CVPR conference
☆ AMUSE: Adaptive Model Updating using a Simulated Environment
Prediction models frequently face the challenge of concept drift, in which the underlying data distribution changes over time, weakening performance. Examples can include models which predict loan default, or those used in healthcare contexts. Typical management strategies involve regular model updates or updates triggered by concept drift detection. However, these simple policies do not necessarily balance the cost of model updating with improved classifier performance. We present AMUSE (Adaptive Model Updating using a Simulated Environment), a novel method leveraging reinforcement learning trained within a simulated data generating environment, to determine update timings for classifiers. The optimal updating policy depends on the current data generating process and ongoing drift process. Our key idea is that we can train an arbitrarily complex model updating policy by creating a training environment in which possible episodes of drift are simulated by a parametric model, which represents expectations of possible drift patterns. As a result, AMUSE proactively recommends updates based on estimated performance improvements, learning a policy that balances maintaining model performance with minimizing update costs. Empirical results confirm the effectiveness of AMUSE in simulated data.
comment: 12 pages, 2 tables. Submitted to AIStats 2025 (under review)
☆ CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Models
In our previous work, we introduced CosyVoice, a multilingual speech synthesis model based on supervised discrete speech tokens. By employing progressive semantic decoding with two popular generative models, language models (LMs) and Flow Matching, CosyVoice demonstrated high prosody naturalness, content consistency, and speaker similarity in speech in-context learning. Recently, significant progress has been made in multi-modal large language models (LLMs), where the response latency and real-time factor of speech synthesis play a crucial role in the interactive experience. Therefore, in this report, we present an improved streaming speech synthesis model, CosyVoice 2, which incorporates comprehensive and systematic optimizations. Specifically, we introduce finite-scalar quantization to improve the codebook utilization of speech tokens. For the text-speech LM, we streamline the model architecture to allow direct use of a pre-trained LLM as the backbone. In addition, we develop a chunk-aware causal flow matching model to support various synthesis scenarios, enabling both streaming and non-streaming synthesis within a single model. By training on a large-scale multilingual dataset, CosyVoice 2 achieves human-parity naturalness, minimal response latency, and virtually lossless synthesis quality in the streaming mode. We invite readers to listen to the demos at https://funaudiollm.github.io/cosyvoice2.
comment: Tech report, work in progress
☆ NetOrchLLM: Mastering Wireless Network Orchestration with Large Language Models
The transition to 6G networks promises unprecedented advancements in wireless communication, with increased data rates, ultra-low latency, and enhanced capacity. However, the complexity of managing and optimizing these next-generation networks presents significant challenges. The advent of large language models (LLMs) has revolutionized various domains by leveraging their sophisticated natural language understanding capabilities. However, the practical application of LLMs in wireless network orchestration and management remains largely unexplored. Existing literature predominantly offers visionary perspectives without concrete implementations, leaving a significant gap in the field. To address this gap, this paper presents NETORCHLLM, a wireless NETwork ORCHestrator LLM framework that uses LLMs to seamlessly orchestrate diverse wireless-specific models from wireless communication communities using their language understanding and generation capabilities. A comprehensive framework is introduced, demonstrating the practical viability of our approach and showcasing how LLMs can be effectively harnessed to optimize dense network operations, manage dynamic environments, and improve overall network performance. NETORCHLLM bridges the theoretical aspirations of prior research with practical, actionable solutions, paving the way for future advancements in integrating generative AI technologies within the wireless communications sector.
☆ Reward Machine Inference for Robotic Manipulation
Learning from Demonstrations (LfD) and Reinforcement Learning (RL) have enabled robot agents to accomplish complex tasks. Reward Machines (RMs) enhance RL's capability to train policies over extended time horizons by structuring high-level task information. In this work, we introduce a novel LfD approach for learning RMs directly from visual demonstrations of robotic manipulation tasks. Unlike previous methods, our approach requires no predefined propositions or prior knowledge of the underlying sparse reward signals. Instead, it jointly learns the RM structure and identifies key high-level events that drive transitions between RM states. We validate our method on vision-based manipulation tasks, showing that the inferred RM accurately captures task structure and enables an RL agent to effectively learn an optimal policy.
☆ HiTZ at VarDial 2025 NorSID: Overcoming Data Scarcity with Language Transfer and Automatic Data Annotation
In this paper we present our submission for the NorSID Shared Task as part of the 2025 VarDial Workshop (Scherrer et al., 2025), consisting of three tasks: Intent Detection, Slot Filling and Dialect Identification, evaluated using data in different dialects of the Norwegian language. For Intent Detection and Slot Filling, we have fine-tuned a multitask model in a cross-lingual setting, to leverage the xSID dataset available in 17 languages. In the case of Dialect Identification, our final submission consists of a model fine-tuned on the provided development set, which has obtained the highest scores within our experiments. Our final results on the test set show that our models do not drop in performance compared to the development set, likely due to the domain-specificity of the dataset and the similar distribution of both subsets. Finally, we also report an in-depth analysis of the provided datasets and their artifacts, as well as other sets of experiments that have been carried out but did not yield the best results. Additionally, we present an analysis on the reasons why some methods have been more successful than others; mainly the impact of the combination of languages and domain-specificity of the training data on the results.
comment: Vardial 2025 NorSID Shared Task
☆ A Survey on Knowledge Graph Structure and Knowledge Graph Embeddings
Knowledge Graphs (KGs) and their machine learning counterpart, Knowledge Graph Embedding Models (KGEMs), have seen ever-increasing use in a wide variety of academic and applied settings. In particular, KGEMs are typically applied to KGs to solve the link prediction task; i.e. to predict new facts in the domain of a KG based on existing, observed facts. While this approach has been shown substantial power in many end-use cases, it remains incompletely characterised in terms of how KGEMs react differently to KG structure. This is of particular concern in light of recent studies showing that KG structure can be a significant source of bias as well as partially determinant of overall KGEM performance. This paper seeks to address this gap in the state-of-the-art. This paper provides, to the authors' knowledge, the first comprehensive survey exploring established relationships of Knowledge Graph Embedding Models and Graph structure in the literature. It is the hope of the authors that this work will inspire further studies in this area, and contribute to a more holistic understanding of KGs, KGEMs, and the link prediction task.
☆ Text2Cypher: Bridging Natural Language and Graph Databases
Knowledge graphs use nodes, relationships, and properties to represent arbitrarily complex data. When stored in a graph database, the Cypher query language enables efficient modeling and querying of knowledge graphs. However, using Cypher requires specialized knowledge, which can present a challenge for non-expert users. Our work Text2Cypher aims to bridge this gap by translating natural language queries into Cypher query language and extending the utility of knowledge graphs to non-technical expert users. While large language models (LLMs) can be used for this purpose, they often struggle to capture complex nuances, resulting in incomplete or incorrect outputs. Fine-tuning LLMs on domain-specific datasets has proven to be a more promising approach, but the limited availability of high-quality, publicly available Text2Cypher datasets makes this challenging. In this work, we show how we combined, cleaned and organized several publicly available datasets into a total of 44,387 instances, enabling effective fine-tuning and evaluation. Models fine-tuned on this dataset showed significant performance gains, with improvements in Google-BLEU and Exact Match scores over baseline models, highlighting the importance of high-quality datasets and fine-tuning in improving Text2Cypher performance.
☆ Enhanced Speech Emotion Recognition with Efficient Channel Attention Guided Deep CNN-BiLSTM Framework
Speech emotion recognition (SER) is crucial for enhancing affective computing and enriching the domain of human-computer interaction. However, the main challenge in SER lies in selecting relevant feature representations from speech signals with lower computational costs. In this paper, we propose a lightweight SER architecture that integrates attention-based local feature blocks (ALFBs) to capture high-level relevant feature vectors from speech signals. We also incorporate a global feature block (GFB) technique to capture sequential, global information and long-term dependencies in speech signals. By aggregating attention-based local and global contextual feature vectors, our model effectively captures the internal correlation between salient features that reflect complex human emotional cues. To evaluate our approach, we extracted four types of spectral features from speech audio samples: mel-frequency cepstral coefficients, mel-spectrogram, root mean square value, and zero-crossing rate. Through a 5-fold cross-validation strategy, we tested the proposed method on five multi-lingual standard benchmark datasets: TESS, RAVDESS, BanglaSER, SUBESCO, and Emo-DB, and obtained a mean accuracy of 99.65%, 94.88%, 98.12%, 97.94%, and 97.19% respectively. The results indicate that our model achieves state-of-the-art (SOTA) performance compared to most existing methods.
comment: 42 pages,10 figures
☆ Class flipping for uplift modeling and Heterogeneous Treatment Effect estimation on imbalanced RCT data
Uplift modeling and Heterogeneous Treatment Effect (HTE) estimation aim at predicting the causal effect of an action, such as a medical treatment or a marketing campaign on a specific individual. In this paper, we focus on data from Randomized Controlled Experiments which guarantee causal interpretation of the outcomes. Class and treatment imbalance are important problems in uplift modeling/HTE, but classical undersampling or oversampling based approaches are hard to apply in this case since they distort the predicted effect. Calibration methods have been proposed in the past, however, they do not guarantee correct predictions. In this work, we propose an approach alternative to undersampling, based on flipping the class value of selected records. We show that the proposed approach does not distort the predicted effect and does not require calibration. The method is especially useful for models based on class variable transformation (modified outcome models). We address those models separately, designing a transformation scheme which guarantees correct predictions and addresses also the problem of treatment imbalance which is especially important for those models. Experiments fully confirm our theoretical results. Additionally, we demonstrate that our method is a viable alternative also for standard classification problems.
☆ Matrix Completion via Residual Spectral Matching
Noisy matrix completion has attracted significant attention due to its applications in recommendation systems, signal processing and image restoration. Most existing works rely on (weighted) least squares methods under various low-rank constraints. However, minimizing the sum of squared residuals is not always efficient, as it may ignore the potential structural information in the residuals.In this study, we propose a novel residual spectral matching criterion that incorporates not only the numerical but also locational information of residuals. This criterion is the first in noisy matrix completion to adopt the perspective of low-rank perturbation of random matrices and exploit the spectral properties of sparse random matrices. We derive optimal statistical properties by analyzing the spectral properties of sparse random matrices and bounding the effects of low-rank perturbations and partial observations. Additionally, we propose algorithms that efficiently approximate solutions by constructing easily computable pseudo-gradients. The iterative process of the proposed algorithms ensures convergence at a rate consistent with the optimal statistical error bound. Our method and algorithms demonstrate improved numerical performance in both simulated and real data examples, particularly in environments with high noise levels.
comment: 23 pages, 6 figures
☆ One Filter to Deploy Them All: Robust Safety for Quadrupedal Navigation in Unknown Environments
As learning-based methods for legged robots rapidly grow in popularity, it is important that we can provide safety assurances efficiently across different controllers and environments. Existing works either rely on a priori knowledge of the environment and safety constraints to ensure system safety or provide assurances for a specific locomotion policy. To address these limitations, we propose an observation-conditioned reachability-based (OCR) safety-filter framework. Our key idea is to use an OCR value network (OCR-VN) that predicts the optimal control-theoretic safety value function for new failure regions and dynamic uncertainty during deployment time. Specifically, the OCR-VN facilitates rapid safety adaptation through two key components: a LiDAR-based input that allows the dynamic construction of safe regions in light of new obstacles and a disturbance estimation module that accounts for dynamics uncertainty in the wild. The predicted safety value function is used to construct an adaptive safety filter that overrides the nominal quadruped controller when necessary to maintain safety. Through simulation studies and hardware experiments on a Unitree Go1 quadruped, we demonstrate that the proposed framework can automatically safeguard a wide range of hierarchical quadruped controllers, adapts to novel environments, and is robust to unmodeled dynamics without a priori access to the controllers or environments - hence, "One Filter to Deploy Them All". The experiment videos can be found on the project website.
comment: Project website: https://sia-lab-git.github.io/One_Filter_to_Deploy_Them_All/
☆ Real-Time Fall Detection Using Smartphone Accelerometers and WiFi Channel State Information
In recent years, as the population ages, falls have increasingly posed a significant threat to the health of the elderly. We propose a real-time fall detection system that integrates the inertial measurement unit (IMU) of a smartphone with optimized Wi-Fi channel state information (CSI) for secondary validation. Initially, the IMU distinguishes falls from routine daily activities with minimal computational demand. Subsequently, the CSI is employed for further assessment, which includes evaluating the individual's post-fall mobility. This methodology not only achieves high accuracy but also reduces energy consumption in the smartphone platform. An Android application developed specifically for the purpose issues an emergency alert if the user experiences a fall and is unable to move. Experimental results indicate that the CSI model, based on convolutional neural networks (CNN), achieves a detection accuracy of 99%, \revised{surpassing comparable IMU-only models, and demonstrating significant resilience in distinguishing between falls and non-fall activities.
☆ Efficient Large-Scale Traffic Forecasting with Transformers: A Spatial Data Management Perspective KDD 2025
Road traffic forecasting is crucial in real-world intelligent transportation scenarios like traffic dispatching and path planning in city management and personal traveling. Spatio-temporal graph neural networks (STGNNs) stand out as the mainstream solution in this task. Nevertheless, the quadratic complexity of remarkable dynamic spatial modeling-based STGNNs has become the bottleneck over large-scale traffic data. From the spatial data management perspective, we present a novel Transformer framework called PatchSTG to efficiently and dynamically model spatial dependencies for large-scale traffic forecasting with interpretability and fidelity. Specifically, we design a novel irregular spatial patching to reduce the number of points involved in the dynamic calculation of Transformer. The irregular spatial patching first utilizes the leaf K-dimensional tree (KDTree) to recursively partition irregularly distributed traffic points into leaf nodes with a small capacity, and then merges leaf nodes belonging to the same subtree into occupancy-equaled and non-overlapped patches through padding and backtracking. Based on the patched data, depth and breadth attention are used interchangeably in the encoder to dynamically learn local and global spatial knowledge from points in a patch and points with the same index of patches. Experimental results on four real world large-scale traffic datasets show that our PatchSTG achieves train speed and memory utilization improvements up to $10\times$ and $4\times$ with the state-of-the-art performance.
comment: Accepted by SIGKDD 2025
☆ GraSP: Simple yet Effective Graph Similarity Predictions AAAI2025
Graph similarity computation (GSC) is to calculate the similarity between one pair of graphs, which is a fundamental problem with fruitful applications in the graph community. In GSC, graph edit distance (GED) and maximum common subgraph (MCS) are two important similarity metrics, both of which are NP-hard to compute. Instead of calculating the exact values, recent solutions resort to leveraging graph neural networks (GNNs) to learn data-driven models for the estimation of GED and MCS. Most of them are built on components involving node-level interactions crossing graphs, which engender vast computation overhead but are of little avail in effectiveness. In the paper, we present GraSP, a simple yet effective GSC approach for GED and MCS prediction. GraSP achieves high result efficacy through several key instruments: enhanced node features via positional encoding and a GNN model augmented by a gating mechanism, residual connections, as well as multi-scale pooling. Theoretically, GraSP can surpass the 1-WL test, indicating its high expressiveness. Empirically, extensive experiments comparing GraSP against 10 competitors on multiple widely adopted benchmark datasets showcase the superiority of GraSP over prior arts in terms of both effectiveness and efficiency. The code is available at https://github.com/HaoranZ99/GraSP.
comment: Accepted by AAAI2025. 13 pages, 14 figures. The code is available at https://github.com/HaoranZ99/GraSP
☆ EP-CFG: Energy-Preserving Classifier-Free Guidance
Classifier-free guidance (CFG) is widely used in diffusion models but often introduces over-contrast and over-saturation artifacts at higher guidance strengths. We present EP-CFG (Energy-Preserving Classifier-Free Guidance), which addresses these issues by preserving the energy distribution of the conditional prediction during the guidance process. Our method simply rescales the energy of the guided output to match that of the conditional prediction at each denoising step, with an optional robust variant for improved artifact suppression. Through experiments, we show that EP-CFG maintains natural image quality and preserves details across guidance strengths while retaining CFG's semantic alignment benefits, all with minimal computational overhead.
☆ What constitutes a Deep Fake? The blurry line between legitimate processing and manipulation under the EU AI Act
When does a digital image resemble reality? The relevance of this question increases as the generation of synthetic images -- so called deep fakes -- becomes increasingly popular. Deep fakes have gained much attention for a number of reasons -- among others, due to their potential to disrupt the political climate. In order to mitigate these threats, the EU AI Act implements specific transparency regulations for generating synthetic content or manipulating existing content. However, the distinction between real and synthetic images is -- even from a computer vision perspective -- far from trivial. We argue that the current definition of deep fakes in the AI act and the corresponding obligations are not sufficiently specified to tackle the challenges posed by deep fakes. By analyzing the life cycle of a digital photo from the camera sensor to the digital editing features, we find that: (1.) Deep fakes are ill-defined in the EU AI Act. The definition leaves too much scope for what a deep fake is. (2.) It is unclear how editing functions like Google's ``best take'' feature can be considered as an exception to transparency obligations. (3.) The exception for substantially edited images raises questions about what constitutes substantial editing of content and whether or not this editing must be perceptible by a natural person. Our results demonstrate that complying with the current AI Act transparency obligations is difficult for providers and deployers. As a consequence of the unclear provisions, there is a risk that exceptions may be either too broad or too limited. We intend our analysis to foster the discussion on what constitutes a deep fake and to raise awareness about the pitfalls in the current AI Act transparency obligations.
comment: Preprint. Accepted at ACM CS&Law '25
☆ Llama 3 Meets MoE: Efficient Upcycling
Scaling large language models (LLMs) significantly improves performance but comes with prohibitive computational costs. Mixture-of-Experts (MoE) models offer an efficient alternative, increasing capacity without a proportional rise in compute requirements. However, training MoE models from scratch poses challenges like overfitting and routing instability. We present an efficient training recipe leveraging pre-trained dense checkpoints, training an 8-Expert Top-2 MoE model from Llama 3-8B with less than $1\%$ of typical pre-training compute. Our approach enhances downstream performance on academic benchmarks, achieving a $\textbf{2%}$ improvement in 0-shot accuracy on MMLU, while reaching a Model FLOPs Utilization (MFU) of $\textbf{46.8%}$ during training using our framework. We also integrate online upcycling in NeMo for seamless use of pre-trained weights, enabling cost-effective development of high-capacity MoE models.
☆ Towards Fair Graph Neural Networks via Graph Counterfactual without Sensitive Attributes ICDE 2025
Graph-structured data is ubiquitous in today's connected world, driving extensive research in graph analysis. Graph Neural Networks (GNNs) have shown great success in this field, leading to growing interest in developing fair GNNs for critical applications. However, most existing fair GNNs focus on statistical fairness notions, which may be insufficient when dealing with statistical anomalies. Hence, motivated by causal theory, there has been growing attention to mitigating root causes of unfairness utilizing graph counterfactuals. Unfortunately, existing methods for generating graph counterfactuals invariably require the sensitive attribute. Nevertheless, in many real-world applications, it is usually infeasible to obtain sensitive attributes due to privacy or legal issues, which challenge existing methods. In this paper, we propose a framework named Fairwos (improving Fairness without sensitive attributes). In particular, we first propose a mechanism to generate pseudo-sensitive attributes to remedy the problem of missing sensitive attributes, and then design a strategy for finding graph counterfactuals from the real dataset. To train fair GNNs, we propose a method to ensure that the embeddings from the original data are consistent with those from the graph counterfactuals, and dynamically adjust the weight of each pseudo-sensitive attribute to balance its contribution to fairness and utility. Furthermore, we theoretically demonstrate that minimizing the relation between these pseudo-sensitive attributes and the prediction can enable the fairness of GNNs. Experimental results on six real-world datasets show that our approach outperforms state-of-the-art methods in balancing utility and fairness.
comment: ICDE 2025
☆ Latent feedback control of distributed systems in multiple scenarios through deep learning-based reduced order models
Continuous monitoring and real-time control of high-dimensional distributed systems are often crucial in applications to ensure a desired physical behavior, without degrading stability and system performances. Traditional feedback control design that relies on full-order models, such as high-dimensional state-space representations or partial differential equations, fails to meet these requirements due to the delay in the control computation, which requires multiple expensive simulations of the physical system. The computational bottleneck is even more severe when considering parametrized systems, as new strategies have to be determined for every new scenario. To address these challenges, we propose a real-time closed-loop control strategy enhanced by nonlinear non-intrusive Deep Learning-based Reduced Order Models (DL-ROMs). Specifically, in the offline phase, (i) full-order state-control pairs are generated for different scenarios through the adjoint method, (ii) the essential features relevant for control design are extracted from the snapshots through a combination of Proper Orthogonal Decomposition (POD) and deep autoencoders, and (iii) the low-dimensional policy bridging latent control and state spaces is approximated with a feedforward neural network. After data generation and neural networks training, the optimal control actions are retrieved in real-time for any observed state and scenario. In addition, the dynamics may be approximated through a cheap surrogate model in order to close the loop at the latent level, thus continuously controlling the system in real-time even when full-order state measurements are missing. The effectiveness of the proposed method, in terms of computational speed, accuracy, and robustness against noisy data, is finally assessed on two different high-dimensional optimal transport problems, one of which also involving an underlying fluid flow.
☆ Predictive Query-based Pipeline for Graph Data
Graphs face challenges when dealing with massive datasets. They are essential tools for modeling interconnected data and often become computationally expensive. Graph embedding techniques, on the other hand, provide an efficient approach. By projecting complex graphs into a lower-dimensional space, these techniques simplify the analysis and processing of large-scale graphs. By transforming graphs into vectors, it simplifies the analysis and processing of large-scale datasets. Several approaches, such as GraphSAGE, Node2Vec, and FastRP, offer efficient methods for generating graph embeddings. By storing embeddings as node properties, it is possible to compare different embedding techniques and evaluate their effectiveness for specific tasks. This flexibilityallows for dynamic updates to embeddings and facilitates experimentation with different approaches. By analyzing these embeddings, one can extract valuable insights into the relationships between nodes and their similarities within the embedding space
☆ Simulating Hard Attention Using Soft Attention
We study conditions under which transformers using soft attention can simulate hard attention, that is, effectively focus all attention on a subset of positions. First, we examine several variants of linear temporal logic, whose formulas have been previously been shown to be computable using hard attention transformers. We demonstrate how soft attention transformers can compute formulas of these logics using unbounded positional embeddings or temperature scaling. Second, we demonstrate how temperature scaling allows softmax transformers to simulate a large subclass of average-hard attention transformers, those that have what we call the uniform-tieless property.
☆ One Node One Model: Featuring the Missing-Half for Graph Clustering
Most existing graph clustering methods primarily focus on exploiting topological structure, often neglecting the ``missing-half" node feature information, especially how these features can enhance clustering performance. This issue is further compounded by the challenges associated with high-dimensional features. Feature selection in graph clustering is particularly difficult because it requires simultaneously discovering clusters and identifying the relevant features for these clusters. To address this gap, we introduce a novel paradigm called ``one node one model", which builds an exclusive model for each node and defines the node label as a combination of predictions for node groups. Specifically, the proposed ``Feature Personalized Graph Clustering (FPGC)" method identifies cluster-relevant features for each node using a squeeze-and-excitation block, integrating these features into each model to form the final representations. Additionally, the concept of feature cross is developed as a data augmentation technique to learn low-order feature interactions. Extensive experimental results demonstrate that FPGC outperforms state-of-the-art clustering methods. Moreover, the plug-and-play nature of our method provides a versatile solution to enhance GNN-based models from a feature perspective.
☆ Analyzing Fairness of Computer Vision and Natural Language Processing Models
Machine learning (ML) algorithms play a crucial role in decision making across diverse fields such as healthcare, finance, education, and law enforcement. Despite their widespread adoption, these systems raise ethical and social concerns due to potential biases and fairness issues. This study focuses on evaluating and improving the fairness of Computer Vision and Natural Language Processing (NLP) models applied to unstructured datasets, emphasizing how biased predictions can reinforce existing systemic inequalities. A publicly available dataset from Kaggle was utilized to simulate a practical scenario for examining fairness in ML workflows. To address and mitigate biases, the study employed two leading fairness libraries: Fairlearn by Microsoft, and AIF360 by IBM. These tools offer comprehensive frameworks for fairness analysis, including metrics evaluation, result visualization, and bias mitigation techniques. The research aims to measure bias levels in ML models, compare the effectiveness of these fairness libraries, and provide actionable recommendations for practitioners. The results demonstrate that each library possesses distinct strengths and limitations in evaluating and mitigating fairness. By systematically analyzing these tools, the study contributes valuable insights to the growing field of ML fairness, offering practical guidance for integrating fairness solutions into real world applications. This research underscores the importance of building more equitable and responsible machine learning systems.
comment: 16 pages, 1 table, 4 figures
☆ TTAQ: Towards Stable Post-training Quantization in Continuous Domain Adaptation
Post-training quantization (PTQ) reduces excessive hardware cost by quantizing full-precision models into lower bit representations on a tiny calibration set, without retraining. Despite the remarkable progress made through recent efforts, traditional PTQ methods typically encounter failure in dynamic and ever-changing real-world scenarios, involving unpredictable data streams and continual domain shifts, which poses greater challenges. In this paper, we propose a novel and stable quantization process for test-time adaptation (TTA), dubbed TTAQ, to address the performance degradation of traditional PTQ in dynamically evolving test domains. To tackle domain shifts in quantizer, TTAQ proposes the Perturbation Error Mitigation (PEM) and Perturbation Consistency Reconstruction (PCR). Specifically, PEM analyzes the error propagation and devises a weight regularization scheme to mitigate the impact of input perturbations. On the other hand, PCR introduces consistency learning to ensure that quantized models provide stable predictions for same sample. Furthermore, we introduce Adaptive Balanced Loss (ABL) to adjust the logits by taking advantage of the frequency and complexity of the class, which can effectively address the class imbalance caused by unpredictable data streams during optimization. Extensive experiments are conducted on multiple datasets with generic TTA methods, proving that TTAQ can outperform existing baselines and encouragingly improve the accuracy of low bit PTQ models in continually changing test domains. For instance, TTAQ decreases the mean error of 2-bit models on ImageNet-C dataset by an impressive 10.1\%.
☆ Analyzing Fairness of Classification Machine Learning Model with Structured Dataset
Machine learning (ML) algorithms have become integral to decision making in various domains, including healthcare, finance, education, and law enforcement. However, concerns about fairness and bias in these systems pose significant ethical and social challenges. This study investigates the fairness of ML models applied to structured datasets in classification tasks, highlighting the potential for biased predictions to perpetuate systemic inequalities. A publicly available dataset from Kaggle was selected for analysis, offering a realistic scenario for evaluating fairness in machine learning workflows. To assess and mitigate biases, three prominent fairness libraries; Fairlearn by Microsoft, AIF360 by IBM, and the What If Tool by Google were employed. These libraries provide robust frameworks for analyzing fairness, offering tools to evaluate metrics, visualize results, and implement bias mitigation strategies. The research aims to assess the extent of bias in the ML models, compare the effectiveness of these libraries, and derive actionable insights for practitioners. The findings reveal that each library has unique strengths and limitations in fairness evaluation and mitigation. By systematically comparing their capabilities, this study contributes to the growing field of ML fairness by providing practical guidance for integrating fairness tools into real world applications. These insights are intended to support the development of more equitable machine learning systems.
comment: 12 pages, 3 tables
☆ Semi-Periodic Activation for Time Series Classification
This paper investigates the lack of research on activation functions for neural network models in time series tasks. It highlights the need to identify essential properties of these activations to improve their effectiveness in specific domains. To this end, the study comprehensively analyzes properties, such as bounded, monotonic, nonlinearity, and periodicity, for activation in time series neural networks. We propose a new activation that maximizes the coverage of these properties, called LeakySineLU. We empirically evaluate the LeakySineLU against commonly used activations in the literature using 112 benchmark datasets for time series classification, obtaining the best average ranking in all comparative scenarios.
☆ Financial Fine-tuning a Large Time Series Model
Large models have shown unprecedented capabilities in natural language processing, image generation, and most recently, time series forecasting. This leads us to ask the question: treating market prices as a time series, can large models be used to predict the market? In this paper, we answer this by evaluating the performance of the latest time series foundation model TimesFM on price prediction. We find that due to the irregular nature of price data, directly applying TimesFM gives unsatisfactory results and propose to fine-tune TimeFM on financial data for the task of price prediction. This is done by continual pre-training of the latest time series foundation model TimesFM on price data containing 100 million time points, spanning a range of financial instruments spanning hourly and daily granularities. The fine-tuned model demonstrates higher price prediction accuracy than the baseline model. We conduct mock trading for our model in various financial markets and show that it outperforms various benchmarks in terms of returns, sharpe ratio, max drawdown and trading cost.
☆ Data-Driven Transfer Learning Framework for Estimating Turning Movement Counts
Urban transportation networks are vital for the efficient movement of people and goods, necessitating effective traffic management and planning. An integral part of traffic management is understanding the turning movement counts (TMCs) at intersections, Accurate TMCs at intersections are crucial for traffic signal control, congestion mitigation, and road safety. In general, TMCs are obtained using physical sensors installed at intersections, but this approach can be cost-prohibitive and technically challenging, especially for cities with extensive road networks. Recent advancements in machine learning and data-driven approaches have offered promising alternatives for estimating TMCs. Traffic patterns can vary significantly across different intersections due to factors such as road geometry, traffic signal settings, and local driver behaviors. This domain discrepancy limits the generalizability and accuracy of machine learning models when applied to new or unseen intersections. In response to these limitations, this research proposes a novel framework leveraging transfer learning (TL) to estimate TMCs at intersections by using traffic controller event-based data, road infrastructure data, and point-of-interest (POI) data. Evaluated on 30 intersections in Tucson, Arizona, the performance of the proposed TL model was compared with eight state-of-the-art regression models and achieved the lowest values in terms of Mean Absolute Error and Root Mean Square Error.
☆ Brain-inspired Chaotic Graph Backpropagation for Large-scale Combinatorial Optimization
Graph neural networks (GNNs) with unsupervised learning can solve large-scale combinatorial optimization problems (COPs) with efficient time complexity, making them versatile for various applications. However, since this method maps the combinatorial optimization problem to the training process of a graph neural network, and the current mainstream backpropagation-based training algorithms are prone to fall into local minima, the optimization performance is still inferior to the current state-of-the-art (SOTA) COP methods. To address this issue, inspired by possibly chaotic dynamics of real brain learning, we introduce a chaotic training algorithm, i.e. chaotic graph backpropagation (CGBP), which introduces a local loss function in GNN that makes the training process not only chaotic but also highly efficient. Different from existing methods, we show that the global ergodicity and pseudo-randomness of such chaotic dynamics enable CGBP to learn each optimal GNN effectively and globally, thus solving the COP efficiently. We have applied CGBP to solve various COPs, such as the maximum independent set, maximum cut, and graph coloring. Results on several large-scale benchmark datasets showcase that CGBP can outperform not only existing GNN algorithms but also SOTA methods. In addition to solving large-scale COPs, CGBP as a universal learning algorithm for GNNs, i.e. as a plug-in unit, can be easily integrated into any existing method for improving the performance.
☆ Financial Sentiment Analysis: Leveraging Actual and Synthetic Data for Supervised Fine-tuning
The Efficient Market Hypothesis (EMH) highlights the essence of financial news in stock price movement. Financial news comes in the form of corporate announcements, news titles, and other forms of digital text. The generation of insights from financial news can be done with sentiment analysis. General-purpose language models are too general for sentiment analysis in finance. Curated labeled data for fine-tuning general-purpose language models are scare, and existing fine-tuned models for sentiment analysis in finance do not capture the maximum context width. We hypothesize that using actual and synthetic data can improve performance. We introduce BertNSP-finance to concatenate shorter financial sentences into longer financial sentences, and finbert-lc to determine sentiment from digital text. The results show improved performance on the accuracy and the f1 score for the financial phrasebank data with $50\%$ and $100\%$ agreement levels.
☆ RLDG: Robotic Generalist Policy Distillation via Reinforcement Learning
Recent advances in robotic foundation models have enabled the development of generalist policies that can adapt to diverse tasks. While these models show impressive flexibility, their performance heavily depends on the quality of their training data. In this work, we propose Reinforcement Learning Distilled Generalists (RLDG), a method that leverages reinforcement learning to generate high-quality training data for finetuning generalist policies. Through extensive real-world experiments on precise manipulation tasks like connector insertion and assembly, we demonstrate that generalist policies trained with RL-generated data consistently outperform those trained with human demonstrations, achieving up to 40% higher success rates while generalizing better to new tasks. We also provide a detailed analysis that reveals this performance gain stems from both optimized action distributions and improved state coverage. Our results suggest that combining task-specific RL with generalist policy distillation offers a promising approach for developing more capable and efficient robotic manipulation systems that maintain the flexibility of foundation models while achieving the performance of specialized controllers. Videos and code can be found on our project website https://generalist-distillation.github.io
☆ LinGen: Towards High-Resolution Minute-Length Text-to-Video Generation with Linear Computational Complexity
Text-to-video generation enhances content creation but is highly computationally intensive: The computational cost of Diffusion Transformers (DiTs) scales quadratically in the number of pixels. This makes minute-length video generation extremely expensive, limiting most existing models to generating videos of only 10-20 seconds length. We propose a Linear-complexity text-to-video Generation (LinGen) framework whose cost scales linearly in the number of pixels. For the first time, LinGen enables high-resolution minute-length video generation on a single GPU without compromising quality. It replaces the computationally-dominant and quadratic-complexity block, self-attention, with a linear-complexity block called MATE, which consists of an MA-branch and a TE-branch. The MA-branch targets short-to-long-range correlations, combining a bidirectional Mamba2 block with our token rearrangement method, Rotary Major Scan, and our review tokens developed for long video generation. The TE-branch is a novel TEmporal Swin Attention block that focuses on temporal correlations between adjacent tokens and medium-range tokens. The MATE block addresses the adjacency preservation issue of Mamba and improves the consistency of generated videos significantly. Experimental results show that LinGen outperforms DiT (with a 75.6% win rate) in video quality with up to 15$\times$ (11.5$\times$) FLOPs (latency) reduction. Furthermore, both automatic metrics and human evaluation demonstrate our LinGen-4B yields comparable video quality to state-of-the-art models (with a 50.5%, 52.1%, 49.1% win rate with respect to Gen-3, LumaLabs, and Kling, respectively). This paves the way to hour-length movie generation and real-time interactive video generation. We provide 68s video generation results and more examples in our project website: https://lineargen.github.io/.
comment: 20 pages, 20 figures
☆ Understand the Effectiveness of Shortcuts through the Lens of DCA
Difference-of-Convex Algorithm (DCA) is a well-known nonconvex optimization algorithm for minimizing a nonconvex function that can be expressed as the difference of two convex ones. Many famous existing optimization algorithms, such as SGD and proximal point methods, can be viewed as special DCAs with specific DC decompositions, making it a powerful framework for optimization. On the other hand, shortcuts are a key architectural feature in modern deep neural networks, facilitating both training and optimization. We showed that the shortcut neural network gradient can be obtained by applying DCA to vanilla neural networks, networks without shortcut connections. Therefore, from the perspective of DCA, we can better understand the effectiveness of networks with shortcuts. Moreover, we proposed a new architecture called NegNet that does not fit the previous interpretation but performs on par with ResNet and can be included in the DCA framework.
☆ Learning Structural Causal Models from Ordering: Identifiable Flow Models AAAI 2025
In this study, we address causal inference when only observational data and a valid causal ordering from the causal graph are available. We introduce a set of flow models that can recover component-wise, invertible transformation of exogenous variables. Our flow-based methods offer flexible model design while maintaining causal consistency regardless of the number of discretization steps. We propose design improvements that enable simultaneous learning of all causal mechanisms and reduce abduction and prediction complexity to linear O(n) relative to the number of layers, independent of the number of causal variables. Empirically, we demonstrate that our method outperforms previous state-of-the-art approaches and delivers consistent performance across a wide range of structural causal models in answering observational, interventional, and counterfactual questions. Additionally, our method achieves a significant reduction in computational time compared to existing diffusion-based techniques, making it practical for large structural causal models.
comment: Accepted at AAAI 2025
☆ Leveraging Programmatically Generated Synthetic Data for Differentially Private Diffusion Training
Programmatically generated synthetic data has been used in differential private training for classification to enhance performance without privacy leakage. However, as the synthetic data is generated from a random process, the distribution of real data and the synthetic data are distinguishable and difficult to transfer. Therefore, the model trained with the synthetic data generates unrealistic random images, raising challenges to adapt the synthetic data for generative models. In this work, we propose DP-SynGen, which leverages programmatically generated synthetic data in diffusion models to address this challenge. By exploiting the three stages of diffusion models(coarse, context, and cleaning) we identify stages where synthetic data can be effectively utilized. We theoretically and empirically verified that cleaning and coarse stages can be trained without private data, replacing them with synthetic data to reduce the privacy budget. The experimental results show that DP-SynGen improves the quality of generative data by mitigating the negative impact of privacy-induced noise on the generation process.
☆ Multivariate Time Series Clustering for Environmental State Characterization of Ground-Based Gravitational-Wave Detectors
Gravitational-wave observatories like LIGO are large-scale, terrestrial instruments housed in infrastructure that spans a multi-kilometer geographic area and which must be actively controlled to maintain operational stability for long observation periods. Despite exquisite seismic isolation, they remain susceptible to seismic noise and other terrestrial disturbances that can couple undesirable vibrations into the instrumental infrastructure, potentially leading to control instabilities or noise artifacts in the detector output. It is, therefore, critical to characterize the seismic state of these observatories to identify a set of temporal patterns that can inform the detector operators in day-to-day monitoring and diagnostics. On a day-to-day basis, the operators monitor several seismically relevant data streams to diagnose operational instabilities and sources of noise using some simple empirically-determined thresholds. It can be untenable for a human operator to monitor multiple data streams in this manual fashion and thus a distillation of these data-streams into a more human-friendly format is sought. In this paper, we present an end-to-end machine learning pipeline for features-based multivariate time series clustering to achieve this goal and to provide actionable insights to the detector operators by correlating found clusters with events of interest in the detector.
comment: 8 pages, 6 figures, Accepted to The 5th International Workshop on Big Data & AI Tools, Methods, and Use Cases for Innovative Scientific Discovery (BTSD 2024)
☆ Precise Antigen-Antibody Structure Predictions Enhance Antibody Development with HelixFold-Multimer
The accurate prediction of antigen-antibody structures is essential for advancing immunology and therapeutic development, as it helps elucidate molecular interactions that underlie immune responses. Despite recent progress with deep learning models like AlphaFold and RoseTTAFold, accurately modeling antigen-antibody complexes remains a challenge due to their unique evolutionary characteristics. HelixFold-Multimer, a specialized model developed for this purpose, builds on the framework of AlphaFold-Multimer and demonstrates improved precision for antigen-antibody structures. HelixFold-Multimer not only surpasses other models in accuracy but also provides essential insights into antibody development, enabling more precise identification of binding sites, improved interaction prediction, and enhanced design of therapeutic antibodies. These advances underscore HelixFold-Multimer's potential in supporting antibody research and therapeutic innovation.
☆ FDM-Bench: A Comprehensive Benchmark for Evaluating Large Language Models in Additive Manufacturing Tasks
Fused Deposition Modeling (FDM) is a widely used additive manufacturing (AM) technique valued for its flexibility and cost-efficiency, with applications in a variety of industries including healthcare and aerospace. Recent developments have made affordable FDM machines accessible and encouraged adoption among diverse users. However, the design, planning, and production process in FDM require specialized interdisciplinary knowledge. Managing the complex parameters and resolving print defects in FDM remain challenging. These technical complexities form the most critical barrier preventing individuals without technical backgrounds and even professional engineers without training in other domains from participating in AM design and manufacturing. Large Language Models (LLMs), with their advanced capabilities in text and code processing, offer the potential for addressing these challenges in FDM. However, existing research on LLM applications in this field is limited, typically focusing on specific use cases without providing comprehensive evaluations across multiple models and tasks. To this end, we introduce FDM-Bench, a benchmark dataset designed to evaluate LLMs on FDM-specific tasks. FDM-Bench enables a thorough assessment by including user queries across various experience levels and G-code samples that represent a range of anomalies. We evaluate two closed-source models (GPT-4o and Claude 3.5 Sonnet) and two open-source models (Llama-3.1-70B and Llama-3.1-405B) on FDM-Bench. A panel of FDM experts assess the models' responses to user queries in detail. Results indicate that closed-source models generally outperform open-source models in G-code anomaly detection, whereas Llama-3.1-405B demonstrates a slight advantage over other models in responding to user queries. These findings underscore FDM-Bench's potential as a foundational tool for advancing research on LLM capabilities in FDM.
☆ Temporal Causal Discovery in Dynamic Bayesian Networks Using Federated Learning
Traditionally, learning the structure of a Dynamic Bayesian Network has been centralized, with all data pooled in one location. However, in real-world scenarios, data are often dispersed among multiple parties (e.g., companies, devices) that aim to collaboratively learn a Dynamic Bayesian Network while preserving their data privacy and security. In this study, we introduce a federated learning approach for estimating the structure of a Dynamic Bayesian Network from data distributed horizontally across different parties. We propose a distributed structure learning method that leverages continuous optimization so that only model parameters are exchanged during optimization. Experimental results on synthetic and real datasets reveal that our method outperforms other state-of-the-art techniques, particularly when there are many clients with limited individual sample sizes.
comment: 23 pages
☆ The Complexity Dynamics of Grokking
We investigate the phenomenon of generalization through the lens of compression. In particular, we study the complexity dynamics of neural networks to explain grokking, where networks suddenly transition from memorizing to generalizing solutions long after over-fitting the training data. To this end we introduce a new measure of intrinsic complexity for neural networks based on the theory of Kolmogorov complexity. Tracking this metric throughout network training, we find a consistent pattern in training dynamics, consisting of a rise and fall in complexity. We demonstrate that this corresponds to memorization followed by generalization. Based on insights from rate--distortion theory and the minimum description length principle, we lay out a principled approach to lossy compression of neural networks, and connect our complexity measure to explicit generalization bounds. Based on a careful analysis of information capacity in neural networks, we propose a new regularization method which encourages networks towards low-rank representations by penalizing their spectral entropy, and find that our regularizer outperforms baselines in total compression of the dataset.
☆ Universal Inceptive GNNs by Eliminating the Smoothness-generalization Dilemma
Graph Neural Networks (GNNs) have demonstrated remarkable success in various domains, such as transaction and social net-works. However, their application is often hindered by the varyinghomophily levels across different orders of neighboring nodes, ne-cessitating separate model designs for homophilic and heterophilicgraphs. In this paper, we aim to develop a unified framework ca-pable of handling neighborhoods of various orders and homophilylevels. Through theoretical exploration, we identify a previouslyoverlooked architectural aspect in multi-hop learning: the cascadedependency, which leads to asmoothness-generalization dilemma.This dilemma significantly affects the learning process, especiallyin the context of high-order neighborhoods and heterophilic graphs.To resolve this issue, we propose an Inceptive Graph Neural Net-work (IGNN), a universal message-passing framework that replacesthe cascade dependency with an inceptive architecture. IGNN pro-vides independent representations for each hop, allowing personal-ized generalization capabilities, and captures neighborhood-wiserelationships to select appropriate receptive fields. Extensive ex-periments show that our IGNN outperforms 23 baseline methods,demonstrating superior performance on both homophilic and het-erophilic graphs, while also scaling efficiently to large graphs.
comment: 12 pages
☆ deepNoC: A deep learning system to assign the number of contributors to a short tandem repeat DNA profile
A common task in forensic biology is to interpret and evaluate short tandem repeat DNA profiles. The first step in these interpretations is to assign a number of contributors to the profiles, a task that is most often performed manually by a scientist using their knowledge of DNA profile behaviour. Studies using constructed DNA profiles have shown that as DNA profiles become more complex, and the number of DNA-donating individuals increases, the ability for scientists to assign the target number. There have been a number of machine learning algorithms developed that seek to assign the number of contributors to a DNA profile, however due to practical limitations in being able to generate DNA profiles in a laboratory, the algorithms have been based on summaries of the available information. In this work we develop an analysis pipeline that simulates the electrophoretic signal of an STR profile, allowing virtually unlimited, pre-labelled training material to be generated. We show that by simulating 100 000 profiles and training a number of contributors estimation tool using a deep neural network architecture (in an algorithm named deepNoC) that a high level of performance is achieved (89% for 1 to 10 contributors). The trained network can then have fine-tuning training performed with only a few hundred profiles in order to achieve the same accuracy within a specific laboratory. We also build into deepNoC secondary outputs that provide a level of explainability to a user of algorithm, and show how they can be displayed in an intuitive manner.
comment: 29 pages, 8 figures
☆ Infinite-dimensional next-generation reservoir computing
Next-generation reservoir computing (NG-RC) has attracted much attention due to its excellent performance in spatio-temporal forecasting of complex systems and its ease of implementation. This paper shows that NG-RC can be encoded as a kernel ridge regression that makes training efficient and feasible even when the space of chosen polynomial features is very large. Additionally, an extension to an infinite number of covariates is possible, which makes the methodology agnostic with respect to the lags into the past that are considered as explanatory factors, as well as with respect to the number of polynomial covariates, an important hyperparameter in traditional NG-RC. We show that this approach has solid theoretical backing and good behavior based on kernel universality properties previously established in the literature. Various numerical illustrations show that these generalizations of NG-RC outperform the traditional approach in several forecasting applications.
comment: 13 pages, 2 figures, 3 tables
☆ Is it the model or the metric -- On robustness measures of deeplearning models
Determining the robustness of deep learning models is an established and ongoing challenge within automated decision-making systems. With the advent and success of techniques that enable advanced deep learning (DL), these models are being used in widespread applications, including high-stake ones like healthcare, education, border-control. Therefore, it is critical to understand the limitations of these models and predict their regions of failures, in order to create the necessary guardrails for their successful and safe deployment. In this work, we revisit robustness, specifically investigating the sufficiency of robust accuracy (RA), within the context of deepfake detection. We present robust ratio (RR) as a complementary metric, that can quantify the changes to the normalized or probability outcomes under input perturbation. We present a comparison of RA and RR and demonstrate that despite similar RA between models, the models show varying RR under different tolerance (perturbation) levels.
comment: Extended abstract at Northern Lights Deep Learning (NLDL) Conference 2025
☆ A Statistical Analysis for Supervised Deep Learning with Exponential Families for Intrinsically Low-dimensional Data
Recent advances have revealed that the rate of convergence of the expected test error in deep supervised learning decays as a function of the intrinsic dimension and not the dimension $d$ of the input space. Existing literature defines this intrinsic dimension as the Minkowski dimension or the manifold dimension of the support of the underlying probability measures, which often results in sub-optimal rates and unrealistic assumptions. In this paper, we consider supervised deep learning when the response given the explanatory variable is distributed according to an exponential family with a $\beta$-H\"older smooth mean function. We consider an entropic notion of the intrinsic data-dimension and demonstrate that with $n$ independent and identically distributed samples, the test error scales as $\tilde{\mathcal{O}}\left(n^{-\frac{2\beta}{2\beta + \bar{d}_{2\beta}(\lambda)}}\right)$, where $\bar{d}_{2\beta}(\lambda)$ is the $2\beta$-entropic dimension of $\lambda$, the distribution of the explanatory variables. This improves on the best-known rates. Furthermore, under the assumption of an upper-bounded density of the explanatory variables, we characterize the rate of convergence as $\tilde{\mathcal{O}}\left( d^{\frac{2\lfloor\beta\rfloor(\beta + d)}{2\beta + d}}n^{-\frac{2\beta}{2\beta + d}}\right)$, establishing that the dependence on $d$ is not exponential but at most polynomial. We also demonstrate that when the explanatory variable has a lower bounded density, this rate in terms of the number of data samples, is nearly optimal for learning the dependence structure for exponential families.
☆ A Novel Methodology in Credit Spread Prediction Based on Ensemble Learning and Feature Selection
The credit spread is a key indicator in bond investments, offering valuable insights for fixed-income investors to devise effective trading strategies. This study proposes a novel credit spread forecasting model leveraging ensemble learning techniques. To enhance predictive accuracy, a feature selection method based on mutual information is incorporated. Empirical results demonstrate that the proposed methodology delivers superior accuracy in credit spread predictions. Additionally, we present a forecast of future credit spread trends using current data, providing actionable insights for investment decision-making.
comment: 7 pages, 5 figures
♻ ☆ Inverse Reinforcement Learning by Estimating Expertise of Demonstrators AAAI
In Imitation Learning (IL), utilizing suboptimal and heterogeneous demonstrations presents a substantial challenge due to the varied nature of real-world data. However, standard IL algorithms consider these datasets as homogeneous, thereby inheriting the deficiencies of suboptimal demonstrators. Previous approaches to this issue rely on impractical assumptions like high-quality data subsets, confidence rankings, or explicit environmental knowledge. This paper introduces IRLEED, Inverse Reinforcement Learning by Estimating Expertise of Demonstrators, a novel framework that overcomes these hurdles without prior knowledge of demonstrator expertise. IRLEED enhances existing Inverse Reinforcement Learning (IRL) algorithms by combining a general model for demonstrator suboptimality to address reward bias and action variance, with a Maximum Entropy IRL framework to efficiently derive the optimal policy from diverse, suboptimal demonstrations. Experiments in both online and offline IL settings, with simulated and human-generated data, demonstrate IRLEED's adaptability and effectiveness, making it a versatile solution for learning from suboptimal demonstrations.
comment: 11 pages, 4 figures, extended version of AAAI publication
♻ ☆ Geometric sparsification in recurrent neural networks
A common technique for ameliorating the computational costs of running large neural models is sparsification, or the pruning of neural connections during training. Sparse models are capable of maintaining the high accuracy of state of the art models, while functioning at the cost of more parsimonious models. The structures which underlie sparse architectures are, however, poorly understood and not consistent between differently trained models and sparsification schemes. In this paper, we propose a new technique for sparsification of recurrent neural nets (RNNs), called moduli regularization, in combination with magnitude pruning. Moduli regularization leverages the dynamical system induced by the recurrent structure to induce a geometric relationship between neurons in the hidden state of the RNN. By making our regularizing term explicitly geometric, we provide the first, to our knowledge, a priori description of the desired sparse architecture of our neural net, as well as explicit end-to-end learning of RNN geometry. We verify the effectiveness of our scheme under diverse conditions, testing in navigation, natural language processing, and addition RNNs. Navigation is a structurally geometric task, for which there are known moduli spaces, and we show that regularization can be used to reach 90% sparsity while maintaining model performance only when coefficients are chosen in accordance with a suitable moduli space. Natural language processing and addition, however, have no known moduli space in which computations are performed. Nevertheless, we show that moduli regularization induces more stable recurrent neural nets, and achieves high fidelity models above 90% sparsity.
comment: 25 pages, 4 figures
♻ ☆ DroidSpeak: KV Cache Sharing for Efficient Multi-LLM Serving
Large Language Models (LLMs) are increasingly employed in complex workflows, where different LLMs and fine-tuned variants collaboratively address complex tasks. However, these systems face significant inefficiencies due to redundant context processing of the shared context. We propose DroidSpeak, a framework that optimizes context sharing between fine-tuned LLMs derived from the same foundational model. DroidSpeak identifies critical layers in the KV cache and selectively recomputes them, enabling effective reuse of intermediate data while maintaining high accuracy. Our approach balances computational efficiency and task fidelity, significantly reducing inference latency and throughput bottlenecks. Experiments on diverse datasets and model pairs demonstrate that DroidSpeak achieves up to 3x higher throughputs and 2.6x faster prefill times with negligible accuracy loss compared to full recomputation.
♻ ☆ MOREL: Enhancing Adversarial Robustness through Multi-Objective Representation Learning
Extensive research has shown that deep neural networks (DNNs) are vulnerable to slight adversarial perturbations$-$small changes to the input data that appear insignificant but cause the model to produce drastically different outputs. In addition to augmenting training data with adversarial examples generated from a specific attack method, most of the current defense strategies necessitate modifying the original model architecture components to improve robustness or performing test-time data purification to handle adversarial attacks. In this work, we demonstrate that strong feature representation learning during training can significantly enhance the original model's robustness. We propose MOREL, a multi-objective feature representation learning approach, encouraging classification models to produce similar features for inputs within the same class, despite perturbations. Our training method involves an embedding space where cosine similarity loss and multi-positive contrastive loss are used to align natural and adversarial features from the model encoder and ensure tight clustering. Concurrently, the classifier is motivated to achieve accurate predictions. Through extensive experiments, we demonstrate that our approach significantly enhances the robustness of DNNs against white-box and black-box adversarial attacks, outperforming other methods that similarly require no architectural changes or test-time data purification. Our code is available at https://github.com/salomonhotegni/MOREL
♻ ☆ On the Power of Adaptive Weighted Aggregation in Heterogeneous Federated Learning and Beyond
Federated averaging (FedAvg) is the most fundamental algorithm in Federated learning (FL). Previous theoretical results assert that FedAvg convergence and generalization degenerate under heterogeneous clients. However, recent empirical results show that FedAvg can perform well in many real-world heterogeneous tasks. These results reveal an inconsistency between FL theory and practice that is not fully explained. In this paper, we show that common heterogeneity measures contribute to this inconsistency based on rigorous convergence analysis. Furthermore, we introduce a new measure \textit{client consensus dynamics} and prove that \textit{FedAvg can effectively handle client heterogeneity when an appropriate aggregation strategy is used}. Building on this theoretical insight, we present a simple and effective FedAvg variant termed FedAWARE. Extensive experiments on three datasets and two modern neural network architectures demonstrate that FedAWARE ensures faster convergence and better generalization in heterogeneous client settings. Moreover, our results show that FedAWARE can significantly enhance the generalization performance of advanced FL algorithms when used as a plug-in module.
♻ ☆ Multi-Feature Fusion and Compressed Bi-LSTM for Memory-Efficient Heartbeat Classification on Wearable Devices
In this article, we present a resource-efficient approach for electrocardiogram (ECG) based heartbeat classification using multi-feature fusion and bidirectional long short-term memory (Bi-LSTM). The dataset comprises five original classes from the MIT-BIH Arrhythmia Database: Normal (N), Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), Premature Ventricular Contraction (PVC), and Paced Beat (PB). Preprocessing methods including the discrete wavelet transform and dual moving average windows are used to reduce noise and artifacts in the raw ECG signal, and extract the main points (PQRST) of the ECG waveform. Multi-feature fusion is achieved by utilizing time intervals and the proposed under-the-curve areas, which are inherently robust against noise, as input features. Simulations demonstrated that incorporating under-the-curve area features improved the classification accuracy for the challenging RBBB and LBBB classes from 31.4\% to 84.3\% for RBBB, and from 69.6\% to 87.0\% for LBBB. Using a Bi-LSTM network, rather than a conventional LSTM network, resulted in higher accuracy (33.8\% vs 21.8\%) with a 28\% reduction in required network parameters for the RBBB class. Multiple neural network models with varying parameter sizes, including tiny (84k), small (150k), medium (478k), and large (1.25M) models, are developed to achieve high accuracy \textit{across all classes}, a more crucial and challenging goal than overall classification accuracy.
♻ ☆ Fair Decentralized Learning
Decentralized learning (DL) is an emerging approach that enables nodes to collaboratively train a machine learning model without sharing raw data. In many application domains, such as healthcare, this approach faces challenges due to the high level of heterogeneity in the training data's feature space. Such feature heterogeneity lowers model utility and negatively impacts fairness, particularly for nodes with under-represented training data. In this paper, we introduce \textsc{Facade}, a clustering-based DL algorithm specifically designed for fair model training when the training data exhibits several distinct features. The challenge of \textsc{Facade} is to assign nodes to clusters, one for each feature, based on the similarity in the features of their local data, without requiring individual nodes to know apriori which cluster they belong to. \textsc{Facade} (1) dynamically assigns nodes to their appropriate clusters over time, and (2) enables nodes to collaboratively train a specialized model for each cluster in a fully decentralized manner. We theoretically prove the convergence of \textsc{Facade}, implement our algorithm, and compare it against three state-of-the-art baselines. Our experimental results on three datasets demonstrate the superiority of our approach in terms of model accuracy and fairness compared to all three competitors. Compared to the best-performing baseline, \textsc{Facade} on the CIFAR-10 dataset also reduces communication costs by 32.3\% to reach a target accuracy when cluster sizes are imbalanced.
comment: To appear in the proceedings of "3rd IEEE Conference on Secure and Trustworthy Machine Learning" (SatML'25)
♻ ☆ Towards the Characterization of Representations Learned via Capsule-based Network Architectures
Capsule Networks (CapsNets) have been re-introduced as a more compact and interpretable alternative to standard deep neural networks. While recent efforts have proved their compression capabilities, to date, their interpretability properties have not been fully assessed. Here, we conduct a systematic and principled study towards assessing the interpretability of these types of networks. Moreover, we pay special attention towards analyzing the level to which part-whole relationships are indeed encoded within the learned representation. Our analysis in the MNIST, SVHN, PASCAL-part and CelebA datasets suggest that the representations encoded in CapsNets might not be as disentangled nor strictly related to parts-whole relationships as is commonly stated in the literature.
comment: This paper consist of 32 pages including 19 figures. This paper concern about interpretation of capsule networks
♻ ☆ SoK: Decentralized AI (DeAI)
The centralization of Artificial Intelligence (AI) poses significant challenges, including single points of failure, inherent biases, data privacy concerns, and scalability issues. These problems are especially prevalent in closed-source large language models (LLMs), where user data is collected and used without transparency. To mitigate these issues, blockchain-based decentralized AI (DeAI) has emerged as a promising solution. DeAI combines the strengths of both blockchain and AI technologies to enhance the transparency, security, decentralization, and trustworthiness of AI systems. However, a comprehensive understanding of state-of-the-art DeAI development, particularly for active industry solutions, is still lacking. In this work, we present a Systematization of Knowledge (SoK) for blockchain-based DeAI solutions. We propose a taxonomy to classify existing DeAI protocols based on the model lifecycle. Based on this taxonomy, we provide a structured way to clarify the landscape of DeAI protocols and identify their similarities and differences. We analyze the functionalities of blockchain in DeAI, investigating how blockchain features contribute to enhancing the security, transparency, and trustworthiness of AI processes, while also ensuring fair incentives for AI data and model contributors. In addition, we identify key insights and research gaps in developing DeAI protocols, highlighting several critical avenues for future research.
comment: This is a Systematization of Knowledge (SoK) for the rapidly evolving field of Decentralized AI (DeAI). We welcome valuable comments, suggestions, and collaboration to further refine and enhance this work. We hope our contribution will help accelerate the advancement of DeAI
♻ ☆ A Gated Residual Kolmogorov-Arnold Networks for Mixtures of Experts
This paper introduces KAMoE, a novel Mixture of Experts (MoE) framework based on Gated Residual Kolmogorov-Arnold Networks (GRKAN). We propose GRKAN as an alternative to the traditional gating function, aiming to enhance efficiency and interpretability in MoE modeling. Through extensive experiments on digital asset markets and real estate valuation, we demonstrate that KAMoE consistently outperforms traditional MoE architectures across various tasks and model types. Our results show that GRKAN exhibits superior performance compared to standard Gating Residual Networks, particularly in LSTM-based models for sequential tasks. We also provide insights into the trade-offs between model complexity and performance gains in MoE and KAMoE architectures.
♻ ☆ Enhanced Low-Dose CT Image Reconstruction by Domain and Task Shifting Gaussian Denoisers
Computed tomography from a low radiation dose (LDCT) is challenging due to high noise in the projection data. Popular approaches for LDCT image reconstruction are two-stage methods, typically consisting of the filtered backprojection (FBP) algorithm followed by a neural network for LDCT image enhancement. Two-stage methods are attractive for their simplicity and potential for computational efficiency, typically requiring only a single FBP and a neural network forward pass for inference. However, the best reconstruction quality is currently achieved by unrolled iterative methods (Learned Primal-Dual and ItNet), which are more complex and thus have a higher computational cost for training and inference. We propose a method combining the simplicity and efficiency of two-stage methods with state-of-the-art reconstruction quality. Our strategy utilizes a neural network pretrained for Gaussian noise removal from natural grayscale images, fine-tuned for LDCT image enhancement. We call this method FBP-DTSGD (Domain and Task Shifted Gaussian Denoisers) as the fine-tuning is a task shift from Gaussian denoising to enhancing LDCT images and a domain shift from natural grayscale to LDCT images. An ablation study with three different pretrained Gaussian denoisers indicates that the performance of FBP-DTSGD does not depend on a specific denoising architecture, suggesting future advancements in Gaussian denoising could benefit the method. The study also shows that pretraining on natural images enhances LDCT reconstruction quality, especially with limited training data. Notably, pretraining involves no additional cost, as existing pretrained models are used. The proposed method currently holds the top mean position in the LoDoPaB-CT challenge.
comment: 13 pages, 4 figures
♻ ☆ IGNITE: Individualized GeNeration of Imputations in Time-series Electronic health records
Electronic Health Records present a valuable modality for driving personalized medicine, where treatment is tailored to fit individual-level differences. For this purpose, many data-driven machine learning and statistical models rely on the wealth of longitudinal EHRs to study patients' physiological and treatment effects. However, longitudinal EHRs tend to be sparse and highly missing, where missingness could also be informative and reflect the underlying patient's health status. Therefore, the success of data-driven models for personalized medicine highly depends on how the EHR data is represented from physiological data, treatments, and the missing values in the data. To this end, we propose a novel deep-learning model that learns the underlying patient dynamics over time across multivariate data to generate personalized realistic values conditioning on an individual's demographic characteristics and treatments. Our proposed model, IGNITE (Individualized GeNeration of Imputations in Time-series Electronic health records), utilises a conditional dual-variational autoencoder augmented with dual-stage attention to generate missing values for an individual. In IGNITE, we further propose a novel individualized missingness mask (IMM), which helps our model generate values based on the individual's observed data and missingness patterns. We further extend the use of IGNITE from imputing missingness to a personalized data synthesizer, where it generates missing EHRs that were never observed prior or even generates new patients for various applications. We validate our model on three large publicly available datasets and show that IGNITE outperforms state-of-the-art approaches in missing data reconstruction and task prediction.
♻ ☆ A Fixed-Point Approach for Causal Generative Modeling
We propose a novel formalism for describing Structural Causal Models (SCMs) as fixed-point problems on causally ordered variables, eliminating the need for Directed Acyclic Graphs (DAGs), and establish the weakest known conditions for their unique recovery given the topological ordering (TO). Based on this, we design a two-stage causal generative model that first infers in a zero-shot manner a valid TO from observations, and then learns the generative SCM on the ordered variables. To infer TOs, we propose to amortize the learning of TOs on synthetically generated datasets by sequentially predicting the leaves of graphs seen during training. To learn SCMs, we design a transformer-based architecture that exploits a new attention mechanism enabling the modeling of causal structures, and show that this parameterization is consistent with our formalism. Finally, we conduct an extensive evaluation of each method individually, and show that when combined, our model outperforms various baselines on generated out-of-distribution problems. The code is available on \href{https://github.com/microsoft/causica/tree/main/research_experiments/fip}{Github}.
♻ ☆ Trained Transformer Classifiers Generalize and Exhibit Benign Overfitting In-Context
Transformers have the capacity to act as supervised learning algorithms: by properly encoding a set of labeled training ("in-context") examples and an unlabeled test example into an input sequence of vectors of the same dimension, the forward pass of the transformer can produce predictions for that unlabeled test example. A line of recent work has shown that when linear transformers are pre-trained on random instances for linear regression tasks, these trained transformers make predictions using an algorithm similar to that of ordinary least squares. In this work, we investigate the behavior of linear transformers trained on random linear classification tasks. Via an analysis of the implicit regularization of gradient descent, we characterize how many pre-training tasks and in-context examples are needed for the trained transformer to generalize well at test-time. We further show that in some settings, these trained transformers can exhibit "benign overfitting in-context": when in-context examples are corrupted by label flipping noise, the transformer memorizes all of its in-context examples (including those with noisy labels) yet still generalizes near-optimally for clean test examples.
comment: 36 pages; added experiments
♻ ☆ Crack-EdgeSAM Self-Prompting Crack Segmentation System for Edge Devices
Structural health monitoring (SHM) is essential for the early detection of infrastructure defects, such as cracks in concrete bridge pier. but often faces challenges in efficiency and accuracy in complex environments. Although the Segment Anything Model (SAM) achieves excellent segmentation performance, its computational demands limit its suitability for real-time applications on edge devices. To address these challenges, this paper proposes Crack-EdgeSAM, a self-prompting crack segmentation system that integrates YOLOv8 for generating prompt boxes and a fine-tuned EdgeSAM model for crack segmentation. To ensure computational efficiency, the method employs ConvLoRA, a Parameter-Efficient Fine-Tuning (PEFT) technique, along with DiceFocalLoss to fine-tune the EdgeSAM model. Our experimental results on public datasets and the climbing robot automatic inspections demonstrate that the system achieves high segmentation accuracy and significantly enhanced inference speed compared to the most recent methods. Notably, the system processes 1024 x 1024 pixels images at 46 FPS on our PC and 8 FPS on Jetson Orin Nano.
♻ ☆ Efficient Sign-Based Optimization: Accelerating Convergence via Variance Reduction
Sign stochastic gradient descent (signSGD) is a communication-efficient method that transmits only the sign of stochastic gradients for parameter updating. Existing literature has demonstrated that signSGD can achieve a convergence rate of $\mathcal{O}(d^{1/2}T^{-1/4})$, where $d$ represents the dimension and $T$ is the iteration number. In this paper, we improve this convergence rate to $\mathcal{O}(d^{1/2}T^{-1/3})$ by introducing the Sign-based Stochastic Variance Reduction (SSVR) method, which employs variance reduction estimators to track gradients and leverages their signs to update. For finite-sum problems, our method can be further enhanced to achieve a convergence rate of $\mathcal{O}(m^{1/4}d^{1/2}T^{-1/2})$, where $m$ denotes the number of component functions. Furthermore, we investigate the heterogeneous majority vote in distributed settings and introduce two novel algorithms that attain improved convergence rates of $\mathcal{O}(d^{1/2}T^{-1/2} + dn^{-1/2})$ and $\mathcal{O}(d^{1/4}T^{-1/4})$ respectively, outperforming the previous results of $\mathcal{O}(dT^{-1/4} + dn^{-1/2})$ and $\mathcal{O}(d^{3/8}T^{-1/8})$, where $n$ represents the number of nodes. Numerical experiments across different tasks validate the effectiveness of our proposed methods.
♻ ☆ A Deterministic Information Bottleneck Method for Clustering Mixed-Type Data
In this paper, we present an information-theoretic method for clustering mixed-type data, that is, data consisting of both continuous and categorical variables. The proposed approach is built on the deterministic variant of the Information Bottleneck algorithm, designed to optimally compress data while preserving its relevant structural information. We evaluate the performance of our method against four well-established clustering techniques for mixed-type data -- KAMILA, K-Prototypes, Factor Analysis for Mixed Data with K-Means, and Partitioning Around Medoids using Gower's dissimilarity -- using both simulated and real-world datasets. The results highlight that the proposed approach offers a competitive alternative to traditional clustering techniques, particularly under specific conditions where heterogeneity in data poses significant challenges.
comment: Submitted to Pattern Recognition
♻ ☆ Sparse Portfolio Selection via Topological Data Analysis based Clustering
This paper uses topological data analysis (TDA) tools and introduces a data-driven clustering-based stock selection strategy tailored for sparse portfolio construction. Our asset selection strategy exploits the topological features of stock price movements to select a subset of topologically similar (different) assets for a sparse index tracking (Markowitz) portfolio. We introduce new distance measures, which serve as an input to the clustering algorithm, on the space of persistence diagrams and landscapes that consider the time component of a time series. We conduct an empirical analysis on the S\&P index from 2009 to 2022, including a study on the COVID-19 data to validate the robustness of our methodology. Our strategy to integrate TDA with the clustering algorithm significantly enhanced the performance of sparse portfolios across various performance measures in diverse market scenarios.
♻ ☆ Clustering Algorithms and RAG Enhancing Semi-Supervised Text Classification with Large LLMs
This paper introduces a novel semi-supervised learning framework specifically designed for text classification tasks, effectively addressing the challenge of vast datasets with limited labeled examples. By integrating multi-level similarity based data augmentation techniques from Retrieval-Augmented Generation (RAG) to Large Language Model (LLM) rewriting and traditional word substitution-we constructed an intelligent augmentation pipeline. This framework innovatively employs the selection of representative landmarks through clustering, which serve as intermediaries in the retrieval and rewriting processes, ensuring that the augmented data maintains a distribution similar to the original dataset. Empirical results show that even in complex text document classification scenarios with over 100 categories, our method achieves state-of-the-art accuracies of 95.41% and 82.43% on the Reuters and Web of Science datasets, respectively. These findings highlight the effectiveness and broad applicability of our semi-supervised learning approach for text classification tasks.
♻ ☆ Unveiling the optimization process of Physics Informed Neural Networks: How accurate and competitive can PINNs be?
This study investigates the potential accuracy boundaries of physics-informed neural networks, contrasting their approach with previous similar works and traditional numerical methods. We find that selecting improved optimization algorithms significantly enhances the accuracy of the results. Simple modifications to the loss function may also improve precision, offering an additional avenue for enhancement. Despite optimization algorithms having a greater impact on convergence than adjustments to the loss function, practical considerations often favor tweaking the latter due to ease of implementation. On a global scale, the integration of an enhanced optimizer and a marginally adjusted loss function enables a reduction in the loss function by several orders of magnitude across diverse physical problems. Consequently, our results obtained using compact networks (typically comprising 2 or 3 layers of 20-30 neurons) achieve accuracies comparable to finite difference schemes employing thousands of grid points. This study encourages the continued advancement of PINNs and associated optimization techniques for broader applications across various fields.
comment: 63 pages, 25 figures. This is the author-accepted manuscript of the paper published in Journal of Computational Physics
♻ ☆ Hidden Biases of End-to-End Driving Datasets CVPR 2024
End-to-end driving systems have made rapid progress, but have so far not been applied to the challenging new CARLA Leaderboard 2.0. Further, while there is a large body of literature on end-to-end architectures and training strategies, the impact of the training dataset is often overlooked. In this work, we make a first attempt at end-to-end driving for Leaderboard 2.0. Instead of investigating architectures, we systematically analyze the training dataset, leading to new insights: (1) Expert style significantly affects downstream policy performance. (2) In complex data sets, the frames should not be weighted on the basis of simplistic criteria such as class frequencies. (3) Instead, estimating whether a frame changes the target labels compared to previous frames can reduce the size of the dataset without removing important information. By incorporating these findings, our model ranks first and second respectively on the map and sensors tracks of the 2024 CARLA Challenge, and sets a new state-of-the-art on the Bench2Drive test routes. Finally, we uncover a design flaw in the current evaluation metrics and propose a modification for future challenges. Our dataset, code, and pre-trained models are publicly available at https://github.com/autonomousvision/carla_garage.
comment: Technical report for the CVPR 2024 Workshop on Foundation Models for Autonomous Systems. Runner-up of the track 'CARLA Autonomous Driving Challenge' in the 2024 Autonomous Grand Challenge (https://opendrivelab.com/challenge2024/)
♻ ☆ Using Low-Discrepancy Points for Data Compression in Machine Learning: An Experimental Comparison
Low-discrepancy points (also called Quasi-Monte Carlo points) are deterministically and cleverly chosen point sets in the unit cube, which provide an approximation of the uniform distribution. We explore two methods based on such low-discrepancy points to reduce large data sets in order to train neural networks. The first one is the method of Dick and Feischl [4], which relies on digital nets and an averaging procedure. Motivated by our experimental findings, we construct a second method, which again uses digital nets, but Voronoi clustering instead of averaging. Both methods are compared to the supercompress approach of [14], which is a variant of the K-means clustering algorithm. The comparison is done in terms of the compression error for different objective functions and the accuracy of the training of a neural network.
♻ ☆ KAGNNs: Kolmogorov-Arnold Networks meet Graph Learning
In recent years, Graph Neural Networks (GNNs) have become the de facto tool for learning node and graph representations. Most GNNs typically consist of a sequence of neighborhood aggregation (a.k.a., message-passing) layers, within which the representation of each node is updated based on those of its neighbors. The most expressive message-passing GNNs can be obtained through the use of the sum aggregator and of MLPs for feature transformation, thanks to their universal approximation capabilities. However, the limitations of MLPs recently motivated the introduction of another family of universal approximators, called Kolmogorov-Arnold Networks (KANs) which rely on a different representation theorem. In this work, we compare the performance of KANs against that of MLPs on graph learning tasks. We evaluate two different implementations of KANs using two distinct base families of functions, namely B-splines and radial basis functions. We perform extensive experiments on node classification, graph classification and graph regression datasets. Our results indicate that KANs are on-par with or better than MLPs on all studied tasks, making them viable alternatives, at the cost of some computational complexity. Code is available at https: //github.com/RomanBresson/KAGNN.
♻ ☆ $\nabla^2$DFT: A Universal Quantum Chemistry Dataset of Drug-Like Molecules and a Benchmark for Neural Network Potentials NeurIPS2024
Methods of computational quantum chemistry provide accurate approximations of molecular properties crucial for computer-aided drug discovery and other areas of chemical science. However, high computational complexity limits the scalability of their applications. Neural network potentials (NNPs) are a promising alternative to quantum chemistry methods, but they require large and diverse datasets for training. This work presents a new dataset and benchmark called $\nabla^2$DFT that is based on the nablaDFT. It contains twice as much molecular structures, three times more conformations, new data types and tasks, and state-of-the-art models. The dataset includes energies, forces, 17 molecular properties, Hamiltonian and overlap matrices, and a wavefunction object. All calculations were performed at the DFT level ($\omega$B97X-D/def2-SVP) for each conformation. Moreover, $\nabla^2$DFT is the first dataset that contains relaxation trajectories for a substantial number of drug-like molecules. We also introduce a novel benchmark for evaluating NNPs in molecular property prediction, Hamiltonian prediction, and conformational optimization tasks. Finally, we propose an extendable framework for training NNPs and implement 10 models within it.
comment: Published as a conference paper at NeurIPS2024 Track on Datasets and Benchmarks (Poster)
♻ ☆ An Information-Theoretic Framework for Out-of-Distribution Generalization with Applications to Stochastic Gradient Langevin Dynamics
We study the Out-of-Distribution (OOD) generalization in machine learning and propose a general framework that establishes information-theoretic generalization bounds. Our framework interpolates freely between Integral Probability Metric (IPM) and $f$-divergence, which naturally recovers some known results (including Wasserstein- and KL-bounds), as well as yields new generalization bounds. Additionally, we show that our framework admits an optimal transport interpretation. When evaluated in two concrete examples, the proposed bounds either strictly improve upon existing bounds in some cases or match the best existing OOD generalization bounds. Moreover, by focusing on $f$-divergence and combining it with the Conditional Mutual Information (CMI) methods, we derive a family of CMI-based generalization bounds, which include the state-of-the-art ICIMI bound as a special instance. Finally, leveraging these findings, we analyze the generalization of the Stochastic Gradient Langevin Dynamics (SGLD) algorithm, showing that our derived generalization bounds outperform existing information-theoretic generalization bounds in certain scenarios.
comment: This work was accepted in part at the 2024 IEEE International Symposium on Information Theory and the 2024 Canadian Workshop on Information Theory. This work was submitted to IEEE Transactions on Information Theory
♻ ☆ Score and Distribution Matching Policy: Advanced Accelerated Visuomotor Policies via Matched Distillation
Visual-motor policy learning has advanced with architectures like diffusion-based policies, known for modeling complex robotic trajectories. However, their prolonged inference times hinder high-frequency control tasks requiring real-time feedback. While consistency distillation (CD) accelerates inference, it introduces errors that compromise action quality. To address these limitations, we propose the Score and Distribution Matching Policy (SDM Policy), which transforms diffusion-based policies into single-step generators through a two-stage optimization process: score matching ensures alignment with true action distributions, and distribution matching minimizes KL divergence for consistency. A dual-teacher mechanism integrates a frozen teacher for stability and an unfrozen teacher for adversarial training, enhancing robustness and alignment with target distributions. Evaluated on a 57-task simulation benchmark, SDM Policy achieves a 6x inference speedup while having state-of-the-art action quality, providing an efficient and reliable framework for high-frequency robotic tasks.
comment: 17 pages
♻ ☆ BSAFusion: A Bidirectional Stepwise Feature Alignment Network for Unaligned Medical Image Fusion AAAI2025
If unaligned multimodal medical images can be simultaneously aligned and fused using a single-stage approach within a unified processing framework, it will not only achieve mutual promotion of dual tasks but also help reduce the complexity of the model. However, the design of this model faces the challenge of incompatible requirements for feature fusion and alignment; specifically, feature alignment requires consistency among corresponding features, whereas feature fusion requires the features to be complementary to each other. To address this challenge, this paper proposes an unaligned medical image fusion method called Bidirectional Stepwise Feature Alignment and Fusion (BSFA-F) strategy. To reduce the negative impact of modality differences on cross-modal feature matching, we incorporate the Modal Discrepancy-Free Feature Representation (MDF-FR) method into BSFA-F. MDF-FR utilizes a Modality Feature Representation Head (MFRH) to integrate the global information of the input image. By injecting the information contained in MFRH of the current image into other modality images, it effectively reduces the impact of modality differences on feature alignment while preserving the complementary information carried by different images. In terms of feature alignment, BSFA-F employs a bidirectional stepwise alignment deformation field prediction strategy based on the path independence of vector displacement between two points. This strategy solves the problem of large spans and inaccurate deformation field prediction in single-step alignment. Finally, Multi-Modal Feature Fusion block achieves the fusion of aligned features. The experimental results across multiple datasets demonstrate the effectiveness of our method. The source code is available at https://github.com/slrl123/BSAFusion.
comment: Accepted by AAAI2025
♻ ☆ Evidential time-to-event prediction with calibrated uncertainty quantification
Time-to-event analysis provides insights into clinical prognosis and treatment recommendations. However, this task is more challenging than standard regression problems due to the presence of censored observations. Additionally, the lack of confidence assessment, model robustness, and prediction calibration raises concerns about the reliability of predictions. To address these challenges, we propose an evidential regression model specifically designed for time-to-event prediction. The proposed model quantifies both epistemic and aleatory uncertainties using Gaussian Random Fuzzy Numbers and belief functions, providing clinicians with uncertainty-aware survival time predictions. The model is trained by minimizing a generalized negative log-likelihood function accounting for data censoring. Experimental evaluations using simulated datasets with different data distributions and censoring conditions, as well as real-world datasets across diverse clinical applications, demonstrate that our model delivers both accurate and reliable performance, outperforming state-of-the-art methods. These results highlight the potential of our approach for enhancing clinical decision-making in survival analysis.
comment: Preprint submitted to International Journal of Approximate Reasoning
♻ ☆ DG-Mamba: Robust and Efficient Dynamic Graph Structure Learning with Selective State Space Models AAAI
Dynamic graphs exhibit intertwined spatio-temporal evolutionary patterns, widely existing in the real world. Nevertheless, the structure incompleteness, noise, and redundancy result in poor robustness for Dynamic Graph Neural Networks (DGNNs). Dynamic Graph Structure Learning (DGSL) offers a promising way to optimize graph structures. However, aside from encountering unacceptable quadratic complexity, it overly relies on heuristic priors, making it hard to discover underlying predictive patterns. How to efficiently refine the dynamic structures, capture intrinsic dependencies, and learn robust representations, remains under-explored. In this work, we propose the novel DG-Mamba, a robust and efficient Dynamic Graph structure learning framework with the Selective State Space Models (Mamba). To accelerate the spatio-temporal structure learning, we propose a kernelized dynamic message-passing operator that reduces the quadratic time complexity to linear. To capture global intrinsic dynamics, we establish the dynamic graph as a self-contained system with State Space Model. By discretizing the system states with the cross-snapshot graph adjacency, we enable the long-distance dependencies capturing with the selective snapshot scan. To endow learned dynamic structures more expressive with informativeness, we propose the self-supervised Principle of Relevant Information for DGSL to regularize the most relevant yet least redundant information, enhancing global robustness. Extensive experiments demonstrate the superiority of the robustness and efficiency of our DG-Mamba compared with the state-of-the-art baselines against adversarial attacks.
comment: Accepted by the Main Technical Track of the 39th Annual AAAI Conference on Artificial Intelligence (AAAI-2025)
♻ ☆ A Geometry-Aware Message Passing Neural Network for Modeling Aerodynamics over Airfoils
Computational modeling of aerodynamics is a key problem in aerospace engineering, often involving flows interacting with solid objects such as airfoils. Deep surrogate models have emerged as purely data-driven approaches that learn direct mappings from simulation conditions to solutions based on either simulation or experimental data. Here, we consider modeling of incompressible flows over solid objects, wherein geometric structures are a key factor in determining aerodynamics. To effectively incorporate geometries, we propose a message passing scheme that efficiently and expressively integrates the airfoil shape with the mesh representation. Under this framework, we first obtain a representation of the geometry in the form of a latent graph on the airfoil surface. We subsequently propagate this representation to all collocation points through message passing on a directed, bipartite graph. We demonstrate that this framework supports efficient training by downsampling the solution mesh while avoiding distribution shifts at test time when evaluated on the full mesh. To enable our model to be able to distinguish between distinct spatial regimes of dynamics relative to the airfoil, we represent mesh points in both a leading edge and trailing edge coordinate system. We further enhance the expressiveness of our coordinate system representations by embedding our hybrid Polar-Cartesian coordinates using sinusoidal and spherical harmonics bases. We additionally find that a change of basis to canonicalize input representations with respect to inlet velocity substantially improves generalization. Altogether, these design choices lead to a purely data-driven machine learning framework known as GeoMPNN, which won the Best Student Submission award at the NeurIPS 2024 ML4CFD Competition, placing 4th overall. Our code is publicly available as part of the AIRS library (https://github.com/divelab/AIRS).
♻ ☆ Solid Waste Detection, Monitoring and Mapping in Remote Sensing Images: A Survey
The detection and characterization of illegal solid waste disposal sites are essential for environmental protection, particularly for mitigating pollution and health hazards. Improperly managed landfills contaminate soil and groundwater via rainwater infiltration, posing threats to both animals and humans. Traditional landfill identification approaches, such as on-site inspections, are time-consuming and expensive. Remote sensing is a cost-effective solution for the identification and monitoring of solid waste disposal sites that enables broad coverage and repeated acquisitions over time. Earth Observation (EO) satellites, equipped with an array of sensors and imaging capabilities, have been providing high-resolution data for several decades. Researchers proposed specialized techniques that leverage remote sensing imagery to perform a range of tasks such as waste site detection, dumping site monitoring, and assessment of suitable locations for new landfills. This review aims to provide a detailed illustration of the most relevant proposals for the detection and monitoring of solid waste sites by describing and comparing the approaches, the implemented techniques, and the employed data. Furthermore, since the data sources are of the utmost importance for developing an effective solid waste detection model, a comprehensive overview of the satellites and publicly available data sets is presented. Finally, this paper identifies the open issues in the state-of-the-art and discusses the relevant research directions for reducing the costs and improving the effectiveness of novel solid waste detection methods.
♻ ☆ Estimating Conditional Average Treatment Effects via Sufficient Representation Learning
Estimating the conditional average treatment effects (CATE) is very important in causal inference and has a wide range of applications across many fields. In the estimation process of CATE, the unconfoundedness assumption is typically required to ensure the identifiability of the regression problems. When estimating CATE using high-dimensional data, there have been many variable selection methods and neural network approaches based on representation learning, while these methods do not provide a way to verify whether the subset of variables after dimensionality reduction or the learned representations still satisfy the unconfoundedness assumption during the estimation process, which can lead to ineffective estimates of the treatment effects. Additionally, these methods typically use data from only the treatment or control group when estimating the regression functions for each group. This paper proposes a novel neural network approach named \textbf{CrossNet} to learn a sufficient representation for the features, based on which we then estimate the CATE, where cross indicates that in estimating the regression functions, we used data from their own group as well as cross-utilized data from another group. Numerical simulations and empirical results demonstrate that our method outperforms the competitive approaches.
♻ ☆ Andes: Defining and Enhancing Quality-of-Experience in LLM-Based Text Streaming Services
Large language models (LLMs) are now at the core of conversational AI services such as real-time translation and chatbots, which provide live user interaction by incrementally streaming text to the user. However, existing LLM serving systems fail to provide good user experience because their optimization metrics are not always aligned with user experience. In this paper, we first introduce and define the notion of Quality-of-Experience (QoE) for text streaming services by considering each user's end-to-end interaction timeline. Based on this, we propose Andes, a QoE-aware LLM serving system that enhances user experience by ensuring that users receive the first token promptly and subsequent tokens at a smooth, digestible pace, even during surge periods. This is enabled by Andes's preemptive request scheduler that dynamically prioritizes requests at the token granularity based on each request's expected QoE gain and GPU resource usage. Our evaluations demonstrate that, compared to state-of-the-art LLM serving systems, Andes improves the average QoE by up to $4.7\times$ given the same GPU resource, or saves up to 61% GPU resources while maintaining the same high QoE.
comment: 16 pages, 21 figures
♻ ☆ Synthetic Potential Outcomes and Causal Mixture Identifiability
Heterogeneous data from multiple populations, sub-groups, or sources is often represented as a ``mixture model'' with a single latent class influencing all of the observed covariates. Heterogeneity can be resolved at multiple levels by grouping populations according to different notions of similarity. This paper proposes grouping with respect to the causal response of an intervention or perturbation on the system. This definition is distinct from previous notions, such as similar covariate values (e.g. clustering) or similar correlations between covariates (e.g. Gaussian mixture models). To solve the problem, we ``synthetically sample'' from a counterfactual distribution using higher-order multi-linear moments of the observable data. To understand how these ``causal mixtures'' fit in with more classical notions, we develop a hierarchy of mixture identifiability.
♻ ☆ Learn and Unlearn in Multilingual LLMs
This paper investigates the propagation of harmful information in multilingual large language models (LLMs) and evaluates the efficacy of various unlearning methods. We demonstrate that fake information, regardless of the language it is in, once introduced into these models through training data, can spread across different languages, compromising the integrity and reliability of the generated content. Our findings reveal that standard unlearning techniques, which typically focus on English data, are insufficient in mitigating the spread of harmful content in multilingual contexts and could inadvertently reinforce harmful content across languages. We show that only by addressing harmful responses in both English and the original language of the harmful data can we effectively eliminate generations for all languages. This underscores the critical need for comprehensive unlearning strategies that consider the multilingual nature of modern LLMs to enhance their safety and reliability across diverse linguistic landscapes.
♻ ☆ Comparing Differentiable and Dynamic Ray Tracing: Introducing the Multipath Lifetime Map
With the increasing presence of dynamic scenarios, such as Vehicle-to-Vehicle communications, radio propagation modeling tools must adapt to the rapidly changing nature of the radio channel. Recently, both Differentiable and Dynamic Ray Tracing frameworks have emerged to address these challenges. However, there is often confusion about how these approaches differ and which one should be used in specific contexts. In this paper, we provide an overview of these two techniques and a comparative analysis against two state-of-the-art tools: 3DSCAT from UniBo and Sionna from NVIDIA. To provide a more precise characterization of the scope of these methods, we introduce a novel simulation-based metric, the Multipath Lifetime Map, which enables the evaluation of spatial and temporal coherence in radio channels only based on the geometrical description of the environment. Finally, our metrics are evaluated on a classic urban street canyon scenario, yielding similar results to those obtained from measurement campaigns.
comment: 5 pages, 5 figures, 1 table, accepted at EuCAP 2025
♻ ☆ MalMixer: Few-Shot Malware Classification with Retrieval-Augmented Semi-Supervised Learning
Recent growth and proliferation of malware has tested practitioners' ability to promptly classify new samples according to malware families. In contrast to labor-intensive reverse engineering efforts, machine learning approaches have demonstrated increased speed and accuracy. However, most existing deep-learning malware family classifiers must be calibrated using a large number of samples that are painstakingly manually analyzed before training. Furthermore, as novel malware samples arise that are beyond the scope of the training set, additional reverse engineering effort must be employed to update the training set. The sheer volume of new samples found in the wild creates substantial pressure on practitioners' ability to reverse engineer enough malware to adequately train modern classifiers. In this paper, we present MalMixer, a malware family classifier using semi-supervised learning that achieves high accuracy with sparse training data. We present a novel domain-knowledge-aware technique for augmenting malware feature representations, enhancing few-shot performance of semi-supervised malware family classification. We show that MalMixer achieves state-of-the-art performance in few-shot malware family classification settings. Our research confirms the feasibility and effectiveness of lightweight, domain-knowledge-aware feature augmentation methods and highlights the capabilities of similar semi-supervised classifiers in addressing malware classification issues.
♻ ☆ Navigating the Future of Federated Recommendation Systems with Foundation Models
In recent years, the integration of federated learning (FL) and recommendation systems (RS), known as Federated Recommendation Systems (FRS), has attracted attention for preserving user privacy by keeping private data on client devices. However, FRS faces inherent limitations such as data heterogeneity and scarcity, due to the privacy requirements of FL and the typical data sparsity issues of RSs. Models like ChatGPT are empowered by the concept of transfer learning and self-supervised learning, so they can be easily applied to the downstream tasks after fine-tuning or prompting. These models, so-called Foundation Models (FM), fouce on understanding the human's intent and perform following their designed roles in the specific tasks, which are widely recognized for producing high-quality content in the image and language domains. Thus, the achievements of FMs inspire the design of FRS and suggest a promising research direction: integrating foundation models to address the above limitations. In this study, we conduct a comprehensive review of FRSs with FMs. Specifically, we: 1) summarise the common approaches of current FRSs and FMs; 2) review the challenges posed by FRSs and FMs; 3) discuss potential future research directions; and 4) introduce some common benchmarks and evaluation metrics in the FRS field. We hope that this position paper provides the necessary background and guidance to explore this interesting and emerging topic.
comment: 20 pages, position paper, survey
♻ ☆ Bootstrapping Heterogeneous Graph Representation Learning via Large Language Models: A Generalized Approach AAAI 2025
Graph representation learning methods are highly effective in handling complex non-Euclidean data by capturing intricate relationships and features within graph structures. However, traditional methods face challenges when dealing with heterogeneous graphs that contain various types of nodes and edges due to the diverse sources and complex nature of the data. Existing Heterogeneous Graph Neural Networks (HGNNs) have shown promising results but require prior knowledge of node and edge types and unified node feature formats, which limits their applicability. Recent advancements in graph representation learning using Large Language Models (LLMs) offer new solutions by integrating LLMs' data processing capabilities, enabling the alignment of various graph representations. Nevertheless, these methods often overlook heterogeneous graph data and require extensive preprocessing. To address these limitations, we propose a novel method that leverages the strengths of both LLM and GNN, allowing for the processing of graph data with any format and type of nodes and edges without the need for type information or special preprocessing. Our method employs LLM to automatically summarize and classify different data formats and types, aligns node features, and uses a specialized GNN for targeted learning, thus obtaining effective graph representations for downstream tasks. Theoretical analysis and experimental validation have demonstrated the effectiveness of our method.
comment: Accepted by AAAI 2025
♻ ☆ Pre-Training Representations of Binary Code Using Contrastive Learning
Binary code analysis and comprehension is critical to applications in reverse engineering and computer security tasks where source code is not available. Unfortunately, unlike source code, binary code lacks semantics and is more difficult for human engineers to understand and analyze. In this paper, we present ContraBin, a contrastive learning technique that integrates source code and comment information along with binaries to create an embedding capable of aiding binary analysis and comprehension tasks. Specifically, we present three components in ContraBin: (1) a primary contrastive learning method for initial pre-training, (2) a simplex interpolation method to integrate source code, comments, and binary code, and (3) an intermediate representation learning algorithm to train a binary code embedding. We further analyze the impact of human-written and synthetic comments on binary code comprehension tasks, revealing a significant performance disparity. While synthetic comments provide substantial benefits, human-written comments are found to introduce noise, even resulting in performance drops compared to using no comments. These findings reshape the narrative around the role of comment types in binary code analysis. We evaluate the effectiveness of ContraBin through four indicative downstream tasks related to binary code: algorithmic functionality classification, function name recovery, code summarization, and reverse engineering. The results show that ContraBin considerably improves performance on all four tasks, measured by accuracy, mean of average precision, and BLEU scores as appropriate. ContraBin is the first language representation model to incorporate source code, binary code, and comments into contrastive code representation learning and is intended to contribute to the field of binary code analysis. The dataset used in this study is available for further research.
♻ ☆ Catch-Up Distillation: You Only Need to Train Once for Accelerating Sampling
Diffusion Probability Models (DPMs) have made impressive advancements in various machine learning domains. However, achieving high-quality synthetic samples typically involves performing a large number of sampling steps, which impedes the possibility of real-time sample synthesis. Traditional accelerated sampling algorithms via knowledge distillation rely on pre-trained model weights and discrete time step scenarios, necessitating additional training sessions to achieve their goals. To address these issues, we propose the Catch-Up Distillation (CUD), which encourages the current moment output of the velocity estimation model ``catch up'' with its previous moment output. Specifically, CUD adjusts the original Ordinary Differential Equation (ODE) training objective to align the current moment output with both the ground truth label and the previous moment output, utilizing Runge-Kutta-based multi-step alignment distillation for precise ODE estimation while preventing asynchronous updates. Furthermore, we investigate the design space for CUDs under continuous time-step scenarios and analyze how to determine the suitable strategies. To demonstrate CUD's effectiveness, we conduct thorough ablation and comparison experiments on CIFAR-10, MNIST, and ImageNet-64. On CIFAR-10, we obtain a FID of 2.80 by sampling in 15 steps under one-session training and the new state-of-the-art FID of 3.37 by sampling in one step with additional training. This latter result necessitated only 620k iterations with a batch size of 128, in contrast to Consistency Distillation, which demanded 2100k iterations with a larger batch size of 256. Our code is released at https://anonymous.4open.science/r/Catch-Up-Distillation-E31F.
♻ ☆ Predicting Many Crystal Properties via an Adaptive Transformer-based Framework
Machine learning has revolutionized many fields, including materials science. However, predicting properties of crystalline materials using machine learning faces challenges in input encoding, output versatility, and interpretability. We introduce CrystalBERT, an adaptable transformer-based framework integrating space group, elemental, and unit cell information. This novel structure can seamlessly combine diverse features and accurately predict various physical properties, including topological properties, superconducting transition temperatures, dielectric constants, and more. CrystalBERT provides insightful interpretations of features influencing target properties. Our results indicate that space group and elemental information are crucial for predicting topological and superconducting properties, underscoring their intricate nature. By incorporating these features, we achieve 91\% accuracy in topological classification, surpassing prior studies and identifying previously misclassified materials. This research demonstrates that integrating diverse material information enhances the prediction of complex material properties, paving the way for more accurate and interpretable machine learning models in materials science.
comment: 38+20 pages, 5+12 figures. The codes are available upon reasonable request
♻ ☆ A Survey of Mamba
As one of the most representative DL techniques, Transformer architecture has empowered numerous advanced models, especially the large language models (LLMs) that comprise billions of parameters, becoming a cornerstone in deep learning. Despite the impressive achievements, Transformers still face inherent limitations, particularly the time-consuming inference resulting from the quadratic computation complexity of attention calculation. Recently, a novel architecture named Mamba, drawing inspiration from classical state space models (SSMs), has emerged as a promising alternative for building foundation models, delivering comparable modeling abilities to Transformers while preserving near-linear scalability concerning sequence length. This has sparked an increasing number of studies actively exploring Mamba's potential to achieve impressive performance across diverse domains. Given such rapid evolution, there is a critical need for a systematic review that consolidates existing Mamba-empowered models, offering a comprehensive understanding of this emerging model architecture. In this survey, we therefore conduct an in-depth investigation of recent Mamba-associated studies, covering three main aspects: the advancements of Mamba-based models, the techniques of adapting Mamba to diverse data, and the applications where Mamba can excel. Specifically, we first review the foundational knowledge of various representative deep learning models and the details of Mamba-1&2 as preliminaries. Then, to showcase the significance of Mamba for AI, we comprehensively review the related studies focusing on Mamba models' architecture design, data adaptability, and applications. Finally, we present a discussion of current limitations and explore various promising research directions to provide deeper insights for future investigations.
♻ ☆ TrialSynth: Generation of Synthetic Sequential Clinical Trial Data
Analyzing data from past clinical trials is part of the ongoing effort to optimize the design, implementation, and execution of new clinical trials and more efficiently bring life-saving interventions to market. While there have been recent advances in the generation of static context synthetic clinical trial data, due to both limited patient availability and constraints imposed by patient privacy needs, the generation of fine-grained synthetic time-sequential clinical trial data has been challenging. Given that patient trajectories over an entire clinical trial are of high importance for optimizing trial design and efforts to prevent harmful adverse events, there is a significant need for the generation of high-fidelity time-sequence clinical trial data. Here we introduce TrialSynth, a Variational Autoencoder (VAE) designed to address the specific challenges of generating synthetic time-sequence clinical trial data. Distinct from related clinical data VAE methods, the core of our method leverages Hawkes Processes (HP), which are particularly well-suited for modeling event-type and time gap prediction needed to capture the structure of sequential clinical trial data. Our experiments demonstrate that TrialSynth surpasses the performance of other comparable methods that can generate sequential clinical trial data at varying levels of fidelity / privacy tradeoff, enabling the generation of highly accurate event sequences across multiple real-world sequential event datasets with small patient source populations. Notably, our empirical findings highlight that TrialSynth not only outperforms existing clinical sequence-generating methods but also produces data with superior utility while empirically preserving patient privacy.
♻ ☆ Accurate and Scalable Estimation of Epistemic Uncertainty for Graph Neural Networks ICLR 2024
While graph neural networks (GNNs) are widely used for node and graph representation learning tasks, the reliability of GNN uncertainty estimates under distribution shifts remains relatively under-explored. Indeed, while post-hoc calibration strategies can be used to improve in-distribution calibration, they need not also improve calibration under distribution shift. However, techniques which produce GNNs with better intrinsic uncertainty estimates are particularly valuable, as they can always be combined with post-hoc strategies later. Therefore, in this work, we propose G-$\Delta$UQ, a novel training framework designed to improve intrinsic GNN uncertainty estimates. Our framework adapts the principle of stochastic data centering to graph data through novel graph anchoring strategies, and is able to support partially stochastic GNNs. While, the prevalent wisdom is that fully stochastic networks are necessary to obtain reliable estimates, we find that the functional diversity induced by our anchoring strategies when sampling hypotheses renders this unnecessary and allows us to support G-$\Delta$UQ on pretrained models. Indeed, through extensive evaluation under covariate, concept and graph size shifts, we show that G-$\Delta$UQ leads to better calibrated GNNs for node and graph classification. Further, it also improves performance on the uncertainty-based tasks of out-of-distribution detection and generalization gap estimation. Overall, our work provides insights into uncertainty estimation for GNNs, and demonstrates the utility of G-$\Delta$UQ in obtaining reliable estimates.
comment: Published at ICLR 2024; Project page: https://pujacomputes.github.io/gduq/
♻ ☆ SMMF: Square-Matricized Momentum Factorization for Memory-Efficient Optimization
We propose SMMF (Square-Matricized Momentum Factorization), a memory-efficient optimizer that reduces the memory requirement of the widely used adaptive learning rate optimizers, such as Adam, by up to 96%. SMMF enables flexible and efficient factorization of an arbitrary rank (shape) of the first and second momentum tensors during optimization, based on the proposed square-matricization and one-time single matrix factorization. From this, it becomes effectively applicable to any rank (shape) of momentum tensors, i.e., bias, matrix, and any rank-d tensors, prevalent in various deep model architectures, such as CNNs (high rank) and Transformers (low rank), in contrast to existing memory-efficient optimizers that applies only to a particular (rank-2) momentum tensor, e.g., linear layers. We conduct a regret bound analysis of SMMF, which shows that it converges similarly to non-memory-efficient adaptive learning rate optimizers, such as AdamNC, providing a theoretical basis for its competitive optimization capability. In our experiment, SMMF takes up to 96% less memory compared to state-of-the-art memory efficient optimizers, e.g., Adafactor, CAME, and SM3, while achieving comparable model performance on various CNN and Transformer tasks.
♻ ☆ Self-test loss functions for learning weak-form operators and gradient flows
The construction of loss functions presents a major challenge in data-driven modeling involving weak-form operators in PDEs and gradient flows, particularly due to the need to select test functions appropriately. We address this challenge by introducing self-test loss functions, which employ test functions that depend on the unknown parameters, specifically for cases where the operator depends linearly on the unknowns. The proposed self-test loss function conserves energy for gradient flows and coincides with the expected log-likelihood ratio for stochastic differential equations. Importantly, it is quadratic, facilitating theoretical analysis of identifiability and well-posedness of the inverse problem, while also leading to efficient parametric or nonparametric regression algorithms. It is computationally simple, requiring only low-order derivatives or even being entirely derivative-free, and numerical experiments demonstrate its robustness against noisy and discrete data.
♻ ☆ Mixture of Online and Offline Experts for Non-stationary Time Series
We consider a general and realistic scenario involving non-stationary time series, consisting of several offline intervals with different distributions within a fixed offline time horizon, and an online interval that continuously receives new samples. For non-stationary time series, the data distribution in the current online interval may have appeared in previous offline intervals. We theoretically explore the feasibility of applying knowledge from offline intervals to the current online interval. To this end, we propose the Mixture of Online and Offline Experts (MOOE). MOOE learns static offline experts from offline intervals and maintains a dynamic online expert for the current online interval. It then adaptively combines the offline and online experts using a meta expert to make predictions for the samples received in the online interval. Specifically, we focus on theoretical analysis, deriving parameter convergence, regret bounds, and generalization error bounds to prove the effectiveness of the algorithm.
♻ ☆ Hierarchical Prompt Decision Transformer: Improving Few-Shot Policy Generalization with Global and Adaptive Guidance
Decision transformers recast reinforcement learning as a conditional sequence generation problem, offering a simple but effective alternative to traditional value or policy-based methods. A recent key development in this area is the integration of prompting in decision transformers to facilitate few-shot policy generalization. However, current methods mainly use static prompt segments to guide rollouts, limiting their ability to provide context-specific guidance. Addressing this, we introduce a hierarchical prompting approach enabled by retrieval augmentation. Our method learns two layers of soft tokens as guiding prompts: (1) global tokens encapsulating task-level information about trajectories, and (2) adaptive tokens that deliver focused, timestep-specific instructions. The adaptive tokens are dynamically retrieved from a curated set of demonstration segments, ensuring context-aware guidance. Experiments across seven benchmark tasks in the MuJoCo and MetaWorld environments demonstrate the proposed approach consistently outperforms all baseline methods, suggesting that hierarchical prompting for decision transformers is an effective strategy to enable few-shot policy generalization.
♻ ☆ Radiology Report Generation via Multi-objective Preference Optimization AAAI 2025
Automatic Radiology Report Generation (RRG) is an important topic for alleviating the substantial workload of radiologists. Existing RRG approaches rely on supervised regression based on different architectures or additional knowledge injection,while the generated report may not align optimally with radiologists' preferences. Especially, since the preferences of radiologists are inherently heterogeneous and multidimensional, e.g., some may prioritize report fluency, while others emphasize clinical accuracy. To address this problem,we propose a new RRG method via Multi-objective Preference Optimization (MPO) to align the pre-trained RRG model with multiple human preferences, which can be formulated by multi-dimensional reward functions and optimized by multi-objective reinforcement learning (RL). Specifically, we use a preference vector to represent the weight of preferences and use it as a condition for the RRG model. Then, a linearly weighed reward is obtained via a dot product between the preference vector and multi-dimensional reward. Next,the RRG model is optimized to align with the preference vector by optimizing such a reward via RL. In the training stage,we randomly sample diverse preference vectors from the preference space and align the model by optimizing the weighted multi-objective rewards, which leads to an optimal policy on the entire preference space. When inference,our model can generate reports aligned with specific preferences without further fine-tuning. Extensive experiments on two public datasets show the proposed method can generate reports that cater to different preferences in a single model and achieve state-of-the-art performance.
comment: Accepted by AAAI 2025
♻ ☆ SLEB: Streamlining LLMs through Redundancy Verification and Elimination of Transformer Blocks ICML 2024
Large language models (LLMs) have proven to be highly effective across various natural language processing tasks. However, their large number of parameters poses significant challenges for practical deployment. Pruning, a technique aimed at reducing the size and complexity of LLMs, offers a potential solution by removing redundant components from the network. Despite the promise of pruning, existing methods often struggle to achieve substantial end-to-end LLM inference speedup. In this paper, we introduce SLEB, a novel approach designed to streamline LLMs by eliminating redundant transformer blocks. We choose the transformer block as the fundamental unit for pruning, because LLMs exhibit block-level redundancy with high similarity between the outputs of neighboring blocks. This choice allows us to effectively enhance the processing speed of LLMs. Our experimental results demonstrate that SLEB outperforms previous LLM pruning methods in accelerating LLM inference while also maintaining superior perplexity and accuracy, making SLEB as a promising technique for enhancing the efficiency of LLMs. The code is available at: https://github.com/jiwonsong-dev/SLEB.
comment: ICML 2024
♻ ☆ Why Does Dropping Edges Usually Outperform Adding Edges in Graph Contrastive Learning? AAAI 2025
Graph contrastive learning (GCL) has been widely used as an effective self-supervised learning method for graph representation learning. However, how to apply adequate and stable graph augmentation to generating proper views for contrastive learning remains an essential problem. Dropping edges is a primary augmentation in GCL while adding edges is not a common method due to its unstable performance. To our best knowledge, there is no theoretical analysis to study why dropping edges usually outperforms adding edges. To answer this question, we introduce a new metric, namely Error Passing Rate (EPR), to quantify how a graph fits the network. Inspired by the theoretical conclusions, we propose a novel GCL algorithm, Error-PAssing-based Graph Contrastive Learning (EPAGCL), which uses both edge adding and edge dropping as its augmentation. To be specific, we generate views by adding and dropping edges according to the weights derived from EPR. Extensive experiments on various real-world datasets are conducted to validate the correctness of our theoretical analysis and the effectiveness of our proposed algorithm.
comment: Accepted by AAAI 2025
♻ ☆ You Can Wash Hands Better: Accurate Daily Handwashing Assessment with Smartwatches
Hand hygiene is among the most effective daily practices for preventing infectious diseases such as influenza, malaria, and skin infections. While professional guidelines emphasize proper handwashing to reduce the risk of viral infections, surveys reveal that adherence to these recommendations remains low. To address this gap, we propose UWash, a wearable solution leveraging smartwatches to evaluate handwashing procedures, aiming to raise awareness and cultivate high-quality handwashing habits. We frame the task of handwashing assessment as an action segmentation problem, similar to those in computer vision, and introduce a simple yet efficient two-stream UNet-like network to achieve this goal. Experiments involving 51 subjects demonstrate that UWash achieves 92.27% accuracy in handwashing gesture recognition, an error of <0.5 seconds in onset/offset detection, and an error of <5 points in gesture scoring under user-dependent settings. The system also performs robustly in user-independent and user-independent-location-independent evaluations. Remarkably, UWash maintains high performance in real-world tests, including evaluations with 10 random passersby at a hospital 9 months later and 10 passersby in an in-the-wild test conducted 2 years later. UWash is the first system to score handwashing quality based on gesture sequences, offering actionable guidance for improving daily hand hygiene. The code and dataset are publicly available at \url{https://github.com/aiotgroup/UWash}.
comment: Under review. 13 pages, 12 figures, 7 tables
♻ ☆ A Comprehensive Framework for Analyzing the Convergence of Adam: Bridging the Gap with SGD
Adaptive Moment Estimation (Adam) is a cornerstone optimization algorithm in deep learning, widely recognized for its flexibility with adaptive learning rates and efficiency in handling large-scale data. However, despite its practical success, the theoretical understanding of Adam's convergence has been constrained by stringent assumptions, such as almost surely bounded stochastic gradients or uniformly bounded gradients, which are more restrictive than those typically required for analyzing stochastic gradient descent (SGD). In this paper, we introduce a novel and comprehensive framework for analyzing the convergence properties of Adam. This framework offers a versatile approach to establishing Adam's convergence. Specifically, we prove that Adam achieves asymptotic (last iterate sense) convergence in both the almost sure sense and the \(L_1\) sense under the relaxed assumptions typically used for SGD, namely \(L\)-smoothness and the ABC inequality. Meanwhile, under the same assumptions, we show that Adam attains non-asymptotic sample complexity bounds similar to those of SGD.
♻ ☆ Robust Dequantization of the Quantum Singular value Transformation and Quantum Machine Learning Algorithms
Several quantum algorithms for linear algebra problems, and in particular quantum machine learning problems, have been "dequantized" in the past few years. These dequantization results typically hold when classical algorithms can access the data via length-squared sampling. In this work we investigate how robust these dequantization results are. We introduce the notion of approximate length-squared sampling, where classical algorithms are only able to sample from a distribution close to the ideal distribution in total variation distance. While quantum algorithms are natively robust against small perturbations, current techniques in dequantization are not. Our main technical contribution is showing how many techniques from randomized linear algebra can be adapted to work under this weaker assumption as well. We then use these techniques to show that the recent low-rank dequantization framework by Chia, Gily\'en, Li, Lin, Tang and Wang (JACM 2022) and the dequantization framework for sparse matrices by Gharibian and Le Gall (STOC 2022), which are both based on the Quantum Singular Value Transformation, can be generalized to the case of approximate length-squared sampling access to the input. We also apply these results to obtain a robust dequantization of many quantum machine learning algorithms, including quantum algorithms for recommendation systems, supervised clustering and low-rank matrix inversion.
comment: 56 pages; v2: minor changes (final journal version)
♻ ☆ Accessible, At-Home Detection of Parkinson's Disease via Multi-task Video Analysis
Limited accessibility to neurological care leads to underdiagnosed Parkinson's Disease (PD), preventing early intervention. Existing AI-based PD detection methods primarily focus on unimodal analysis of motor or speech tasks, overlooking the multifaceted nature of the disease. To address this, we introduce a large-scale, multi-task video dataset consisting of 1102 sessions (each containing videos of finger tapping, facial expression, and speech tasks captured via webcam) from 845 participants (272 with PD). We propose a novel Uncertainty-calibrated Fusion Network (UFNet) that leverages this multimodal data to enhance diagnostic accuracy. UFNet employs independent task-specific networks, trained with Monte Carlo Dropout for uncertainty quantification, followed by self-attended fusion of features, with attention weights dynamically adjusted based on task-specific uncertainties. To ensure patient-centered evaluation, the participants were randomly split into three sets: 60% for training, 20% for model selection, and 20% for final performance evaluation. UFNet significantly outperformed single-task models in terms of accuracy, area under the ROC curve (AUROC), and sensitivity while maintaining non-inferior specificity. Withholding uncertain predictions further boosted the performance, achieving 88.0+-0.3%$ accuracy, 93.0+-0.2% AUROC, 79.3+-0.9% sensitivity, and 92.6+-0.3% specificity, at the expense of not being able to predict for 2.3+-0.3% data (+- denotes 95% confidence interval). Further analysis suggests that the trained model does not exhibit any detectable bias across sex and ethnic subgroups and is most effective for individuals aged between 50 and 80. Requiring only a webcam and microphone, our approach facilitates accessible home-based PD screening, especially in regions with limited healthcare resources.
♻ ☆ Optimal Multitask Linear Regression and Contextual Bandits under Sparse Heterogeneity
Large and complex datasets are often collected from several, possibly heterogeneous sources. Multitask learning methods improve efficiency by leveraging commonalities across datasets while accounting for possible differences among them. Here, we study multitask linear regression and contextual bandits under sparse heterogeneity, where the source/task-associated parameters are equal to a global parameter plus a sparse task-specific term. We propose a novel two-stage estimator called MOLAR that leverages this structure by first constructing a covariate-wise weighted median of the task-wise linear regression estimates and then shrinking the task-wise estimates towards the weighted median. Compared to task-wise least squares estimates, MOLAR improves the dependence of the estimation error on the data dimension. Extensions of MOLAR to generalized linear models and constructing confidence intervals are discussed in the paper. We then apply MOLAR to develop methods for sparsely heterogeneous multitask contextual bandits, obtaining improved regret guarantees over single-task bandit methods. We further show that our methods are minimax optimal by providing a number of lower bounds. Finally, we support the efficiency of our methods by performing experiments on both synthetic data and the PISA dataset on student educational outcomes from heterogeneous countries.
comment: Journal of the American Statistical Association, 2024
♻ ☆ WormKAN: Are KAN Effective for Identifying and Tracking Concept Drift in Time Series?
Dynamic concepts in time series are crucial for understanding complex systems such as financial markets, healthcare, and online activity logs. These concepts help reveal structures and behaviors in sequential data for better decision-making and forecasting. However, existing models often struggle to detect and track concept drift due to limitations in interpretability and adaptability. To address this challenge, inspired by the flexibility of the recent Kolmogorov-Arnold Network (KAN), we propose WormKAN, a concept-aware KAN-based model to address concept drift in co-evolving time series. WormKAN consists of three key components: Patch Normalization, Temporal Representation Module, and Concept Dynamics. Patch normalization processes co-evolving time series into patches, treating them as fundamental modeling units to capture local dependencies while ensuring consistent scaling. The temporal representation module learns robust latent representations by leveraging a KAN-based autoencoder, complemented by a smoothness constraint, to uncover inter-patch correlations. Concept dynamics identifies and tracks dynamic transitions, revealing structural shifts in the time series through concept identification and drift detection. These transitions, akin to passing through a \textit{wormhole}, are identified by abrupt changes in the latent space. Experiments show that KAN and KAN-based models (WormKAN) effectively segment time series into meaningful concepts, enhancing the identification and tracking of concept drift.
Artificial Intelligence 121
☆ GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Prediction
3D occupancy prediction is important for autonomous driving due to its comprehensive perception of the surroundings. To incorporate sequential inputs, most existing methods fuse representations from previous frames to infer the current 3D occupancy. However, they fail to consider the continuity of driving scenarios and ignore the strong prior provided by the evolution of 3D scenes (e.g., only dynamic objects move). In this paper, we propose a world-model-based framework to exploit the scene evolution for perception. We reformulate 3D occupancy prediction as a 4D occupancy forecasting problem conditioned on the current sensor input. We decompose the scene evolution into three factors: 1) ego motion alignment of static scenes; 2) local movements of dynamic objects; and 3) completion of newly-observed scenes. We then employ a Gaussian world model (GaussianWorld) to explicitly exploit these priors and infer the scene evolution in the 3D Gaussian space considering the current RGB observation. We evaluate the effectiveness of our framework on the widely used nuScenes dataset. Our GaussianWorld improves the performance of the single-frame counterpart by over 2% in mIoU without introducing additional computations. Code: https://github.com/zuosc19/GaussianWorld.
comment: Code is available at: https://github.com/zuosc19/GaussianWorld
☆ GaussianAD: Gaussian-Centric End-to-End Autonomous Driving
Vision-based autonomous driving shows great potential due to its satisfactory performance and low costs. Most existing methods adopt dense representations (e.g., bird's eye view) or sparse representations (e.g., instance boxes) for decision-making, which suffer from the trade-off between comprehensiveness and efficiency. This paper explores a Gaussian-centric end-to-end autonomous driving (GaussianAD) framework and exploits 3D semantic Gaussians to extensively yet sparsely describe the scene. We initialize the scene with uniform 3D Gaussians and use surrounding-view images to progressively refine them to obtain the 3D Gaussian scene representation. We then use sparse convolutions to efficiently perform 3D perception (e.g., 3D detection, semantic map construction). We predict 3D flows for the Gaussians with dynamic semantics and plan the ego trajectory accordingly with an objective of future scene forecasting. Our GaussianAD can be trained in an end-to-end manner with optional perception labels when available. Extensive experiments on the widely used nuScenes dataset verify the effectiveness of our end-to-end GaussianAD on various tasks including motion planning, 3D occupancy prediction, and 4D occupancy forecasting. Code: https://github.com/wzzheng/GaussianAD.
comment: Code is available at: https://github.com/wzzheng/GaussianAD
☆ Apollo: An Exploration of Video Understanding in Large Multimodal Models
Despite the rapid integration of video perception capabilities into Large Multimodal Models (LMMs), the underlying mechanisms driving their video understanding remain poorly understood. Consequently, many design decisions in this domain are made without proper justification or analysis. The high computational cost of training and evaluating such models, coupled with limited open research, hinders the development of video-LMMs. To address this, we present a comprehensive study that helps uncover what effectively drives video understanding in LMMs. We begin by critically examining the primary contributors to the high computational requirements associated with video-LMM research and discover Scaling Consistency, wherein design and training decisions made on smaller models and datasets (up to a critical size) effectively transfer to larger models. Leveraging these insights, we explored many video-specific aspects of video-LMMs, including video sampling, architectures, data composition, training schedules, and more. For example, we demonstrated that fps sampling during training is vastly preferable to uniform frame sampling and which vision encoders are the best for video representation. Guided by these findings, we introduce Apollo, a state-of-the-art family of LMMs that achieve superior performance across different model sizes. Our models can perceive hour-long videos efficiently, with Apollo-3B outperforming most existing $7$B models with an impressive 55.1 on LongVideoBench. Apollo-7B is state-of-the-art compared to 7B LMMs with a 70.9 on MLVU, and 63.3 on Video-MME.
comment: https://apollo-lmms.github.io
☆ A Library for Learning Neural Operators
We present NeuralOperator, an open-source Python library for operator learning. Neural operators generalize neural networks to maps between function spaces instead of finite-dimensional Euclidean spaces. They can be trained and inferenced on input and output functions given at various discretizations, satisfying a discretization convergence properties. Built on top of PyTorch, NeuralOperator provides all the tools for training and deploying neural operator models, as well as developing new ones, in a high-quality, tested, open-source package. It combines cutting-edge models and customizability with a gentle learning curve and simple user interface for newcomers.
☆ A dual contrastive framework
In current multimodal tasks, models typically freeze the encoder and decoder while adapting intermediate layers to task-specific goals, such as region captioning. Region-level visual understanding presents significant challenges for large-scale vision-language models. While limited spatial awareness is a known issue, coarse-grained pretraining, in particular, exacerbates the difficulty of optimizing latent representations for effective encoder-decoder alignment. We propose AlignCap, a framework designed to enhance region-level understanding through fine-grained alignment of latent spaces. Our approach introduces a novel latent feature refinement module that enhances conditioned latent space representations to improve region-level captioning performance. We also propose an innovative alignment strategy, the semantic space alignment module, which boosts the quality of multimodal representations. Additionally, we incorporate contrastive learning in a novel manner within both modules to further enhance region-level captioning performance. To address spatial limitations, we employ a General Object Detection (GOD) method as a data preprocessing pipeline that enhances spatial reasoning at the regional level. Extensive experiments demonstrate that our approach significantly improves region-level captioning performance across various tasks
☆ COMET: Benchmark for Comprehensive Biological Multi-omics Evaluation Tasks and Language Models
As key elements within the central dogma, DNA, RNA, and proteins play crucial roles in maintaining life by guaranteeing accurate genetic expression and implementation. Although research on these molecules has profoundly impacted fields like medicine, agriculture, and industry, the diversity of machine learning approaches-from traditional statistical methods to deep learning models and large language models-poses challenges for researchers in choosing the most suitable models for specific tasks, especially for cross-omics and multi-omics tasks due to the lack of comprehensive benchmarks. To address this, we introduce the first comprehensive multi-omics benchmark COMET (Benchmark for Biological COmprehensive Multi-omics Evaluation Tasks and Language Models), designed to evaluate models across single-omics, cross-omics, and multi-omics tasks. First, we curate and develop a diverse collection of downstream tasks and datasets covering key structural and functional aspects in DNA, RNA, and proteins, including tasks that span multiple omics levels. Then, we evaluate existing foundational language models for DNA, RNA, and proteins, as well as the newly proposed multi-omics method, offering valuable insights into their performance in integrating and analyzing data from different biological modalities. This benchmark aims to define critical issues in multi-omics research and guide future directions, ultimately promoting advancements in understanding biological processes through integrated and different omics data analysis.
☆ TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies
Although large vision-language-action (VLA) models pretrained on extensive robot datasets offer promising generalist policies for robotic learning, they still struggle with spatial-temporal dynamics in interactive robotics, making them less effective in handling complex tasks, such as manipulation. In this work, we introduce visual trace prompting, a simple yet effective approach to facilitate VLA models' spatial-temporal awareness for action prediction by encoding state-action trajectories visually. We develop a new TraceVLA model by finetuning OpenVLA on our own collected dataset of 150K robot manipulation trajectories using visual trace prompting. Evaluations of TraceVLA across 137 configurations in SimplerEnv and 4 tasks on a physical WidowX robot demonstrate state-of-the-art performance, outperforming OpenVLA by 10% on SimplerEnv and 3.5x on real-robot tasks and exhibiting robust generalization across diverse embodiments and scenarios. To further validate the effectiveness and generality of our method, we present a compact VLA model based on 4B Phi-3-Vision, pretrained on the Open-X-Embodiment and finetuned on our dataset, rivals the 7B OpenVLA baseline while significantly improving inference efficiency.
☆ Iris: Breaking GUI Complexity with Adaptive Focus and Self-Refining
Digital agents are increasingly employed to automate tasks in interactive digital environments such as web pages, software applications, and operating systems. While text-based agents built on Large Language Models (LLMs) often require frequent updates due to platform-specific APIs, visual agents leveraging Multimodal Large Language Models (MLLMs) offer enhanced adaptability by interacting directly with Graphical User Interfaces (GUIs). However, these agents face significant challenges in visual perception, particularly when handling high-resolution, visually complex digital environments. This paper introduces Iris, a foundational visual agent that addresses these challenges through two key innovations: Information-Sensitive Cropping (ISC) and Self-Refining Dual Learning (SRDL). ISC dynamically identifies and prioritizes visually dense regions using a edge detection algorithm, enabling efficient processing by allocating more computational resources to areas with higher information density. SRDL enhances the agent's ability to handle complex tasks by leveraging a dual-learning loop, where improvements in referring (describing UI elements) reinforce grounding (locating elements) and vice versa, all without requiring additional annotated data. Empirical evaluations demonstrate that Iris achieves state-of-the-art performance across multiple benchmarks with only 850K GUI annotations, outperforming methods using 10x more training data. These improvements further translate to significant gains in both web and OS agent downstream tasks.
☆ Generative AI in Medicine
The increased capabilities of generative AI have dramatically expanded its possible use cases in medicine. We provide a comprehensive overview of generative AI use cases for clinicians, patients, clinical trial organizers, researchers, and trainees. We then discuss the many challenges -- including maintaining privacy and security, improving transparency and interpretability, upholding equity, and rigorously evaluating models -- which must be overcome to realize this potential, and the open research directions they give rise to.
comment: To appear in the Annual Review of Biomedical Data Science, August 2025
☆ AdvPrefix: An Objective for Nuanced LLM Jailbreaks
Many jailbreak attacks on large language models (LLMs) rely on a common objective: making the model respond with the prefix "Sure, here is (harmful request)". While straightforward, this objective has two limitations: limited control over model behaviors, often resulting in incomplete or unrealistic responses, and a rigid format that hinders optimization. To address these limitations, we introduce AdvPrefix, a new prefix-forcing objective that enables more nuanced control over model behavior while being easy to optimize. Our objective leverages model-dependent prefixes, automatically selected based on two criteria: high prefilling attack success rates and low negative log-likelihood. It can further simplify optimization by using multiple prefixes for a single user request. AdvPrefix can integrate seamlessly into existing jailbreak attacks to improve their performance for free. For example, simply replacing GCG attack's target prefixes with ours on Llama-3 improves nuanced attack success rates from 14% to 80%, suggesting that current alignment struggles to generalize to unseen prefixes. Our work demonstrates the importance of jailbreak objectives in achieving nuanced jailbreaks.
☆ MeshA*: Efficient Path Planing With Motion Primitives
We study a path planning problem where the possible move actions are represented as a finite set of motion primitives aligned with the grid representation of the environment. That is, each primitive corresponds to a short kinodynamically-feasible motion of an agent and is represented as a sequence of the swept cells of a grid. Typically heuristic search, i.e. A*, is conducted over the lattice induced by these primitives (lattice-based planning) to find a path. However due to the large branching factor such search may be inefficient in practice. To this end we suggest a novel technique rooted in the idea of searching over the grid cells (as in vanilla A*) simultaneously fitting the possible sequences of the motion primitives into these cells. The resultant algorithm, MeshA*, provably preserves the guarantees on completeness and optimality, on the one hand, and is shown to notably outperform conventional lattice-based planning (x1.5 decrease in the runtime), on the other hand. Moreover, we suggest an additional pruning technique that additionally decreases the search space of MeshA*. The resultant planner is combined with the regular A* to retain completeness and is shown to further increase the search performance at the cost of negligible decrease of the solution quality.
☆ BrushEdit: All-In-One Image Inpainting and Editing
Image editing has advanced significantly with the development of diffusion models using both inversion-based and instruction-based methods. However, current inversion-based approaches struggle with big modifications (e.g., adding or removing objects) due to the structured nature of inversion noise, which hinders substantial changes. Meanwhile, instruction-based methods often constrain users to black-box operations, limiting direct interaction for specifying editing regions and intensity. To address these limitations, we propose BrushEdit, a novel inpainting-based instruction-guided image editing paradigm, which leverages multimodal large language models (MLLMs) and image inpainting models to enable autonomous, user-friendly, and interactive free-form instruction editing. Specifically, we devise a system enabling free-form instruction editing by integrating MLLMs and a dual-branch image inpainting model in an agent-cooperative framework to perform editing category classification, main object identification, mask acquisition, and editing area inpainting. Extensive experiments show that our framework effectively combines MLLMs and inpainting models, achieving superior performance across seven metrics including mask region preservation and editing effect coherence.
comment: WebPage available at https://liyaowei-stu.github.io/project/BrushEdit/
☆ Interlocking-free Selective Rationalization Through Genetic-based Learning
A popular end-to-end architecture for selective rationalization is the select-then-predict pipeline, comprising a generator to extract highlights fed to a predictor. Such a cooperative system suffers from suboptimal equilibrium minima due to the dominance of one of the two modules, a phenomenon known as interlocking. While several contributions aimed at addressing interlocking, they only mitigate its effect, often by introducing feature-based heuristics, sampling, and ad-hoc regularizations. We present GenSPP, the first interlocking-free architecture for selective rationalization that does not require any learning overhead, as the above-mentioned. GenSPP avoids interlocking by performing disjoint training of the generator and predictor via genetic global search. Experiments on a synthetic and a real-world benchmark show that our model outperforms several state-of-the-art competitors.
☆ DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding
We present DeepSeek-VL2, an advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL, through two key major upgrades. For the vision component, we incorporate a dynamic tiling vision encoding strategy designed for processing high-resolution images with different aspect ratios. For the language component, we leverage DeepSeekMoE models with the Multi-head Latent Attention mechanism, which compresses Key-Value cache into latent vectors, to enable efficient inference and high throughput. Trained on an improved vision-language dataset, DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition, document/table/chart understanding, and visual grounding. Our model series is composed of three variants: DeepSeek-VL2-Tiny, DeepSeek-VL2-Small and DeepSeek-VL2, with 1.0B, 2.8B and 4.5B activated parameters respectively. DeepSeek-VL2 achieves competitive or state-of-the-art performance with similar or fewer activated parameters compared to existing open-source dense and MoE-based models. Codes and pre-trained models are publicly accessible at https://github.com/deepseek-ai/DeepSeek-VL2.
☆ Still "Talking About Large Language Models": Some Clarifications
My paper "Talking About Large Language Models" has more than once been interpreted as advocating a reductionist stance towards large language models. But the paper was not intended that way, and I do not endorse such positions. This short note situates the paper in the context of a larger philosophical project that is concerned with the (mis)use of words rather than metaphysics, in the spirit of Wittgenstein's later writing.
☆ Envisioning National Resources for Artificial Intelligence Research: NSF Workshop Report
This is a report of an NSF workshop titled "Envisioning National Resources for Artificial Intelligence Research" held in Alexandria, Virginia, in May 2024. The workshop aimed to identify initial challenges and opportunities for national resources for AI research (e.g., compute, data, models, etc.) and to facilitate planning for the envisioned National AI Research Resource. Participants included AI and cyberinfrastructure (CI) experts. The report outlines significant findings and identifies needs and recommendations from the workshop.
☆ Trustworthy and Explainable Decision-Making for Workforce allocation
In industrial contexts, effective workforce allocation is crucial for operational efficiency. This paper presents an ongoing project focused on developing a decision-making tool designed for workforce allocation, emphasising the explainability to enhance its trustworthiness. Our objective is to create a system that not only optimises the allocation of teams to scheduled tasks but also provides clear, understandable explanations for its decisions, particularly in cases where the problem is infeasible. By incorporating human-in-the-loop mechanisms, the tool aims to enhance user trust and facilitate interactive conflict resolution. We implemented our approach on a prototype tool/digital demonstrator intended to be evaluated on a real industrial scenario both in terms of performance and user acceptability.
comment: Accepted for presentation at PTHG-24: The Seventh Workshop on Progress Towards the Holy Grail, part of the 30th International Conference on Principles and Practice of Constraint Programming. For more details, visit the workshop webpage: https://freuder.wordpress.com/progress-towards-the-holy-grail-workshops/pthg-24-the-seventh-workshop-on-progress-towards-the-holy-grail/
☆ Cultural Evolution of Cooperation among LLM Agents
Large language models (LLMs) provide a compelling foundation for building generally-capable AI agents. These agents may soon be deployed at scale in the real world, representing the interests of individual humans (e.g., AI assistants) or groups of humans (e.g., AI-accelerated corporations). At present, relatively little is known about the dynamics of multiple LLM agents interacting over many generations of iterative deployment. In this paper, we examine whether a "society" of LLM agents can learn mutually beneficial social norms in the face of incentives to defect, a distinctive feature of human sociality that is arguably crucial to the success of civilization. In particular, we study the evolution of indirect reciprocity across generations of LLM agents playing a classic iterated Donor Game in which agents can observe the recent behavior of their peers. We find that the evolution of cooperation differs markedly across base models, with societies of Claude 3.5 Sonnet agents achieving significantly higher average scores than Gemini 1.5 Flash, which, in turn, outperforms GPT-4o. Further, Claude 3.5 Sonnet can make use of an additional mechanism for costly punishment to achieve yet higher scores, while Gemini 1.5 Flash and GPT-4o fail to do so. For each model class, we also observe variation in emergent behavior across random seeds, suggesting an understudied sensitive dependence on initial conditions. We suggest that our evaluation regime could inspire an inexpensive and informative new class of LLM benchmarks, focussed on the implications of LLM agent deployment for the cooperative infrastructure of society.
comment: 15 pages, 6 figures
☆ Does Multiple Choice Have a Future in the Age of Generative AI? A Posttest-only RCT
The role of multiple-choice questions (MCQs) as effective learning tools has been debated in past research. While MCQs are widely used due to their ease in grading, open response questions are increasingly used for instruction, given advances in large language models (LLMs) for automated grading. This study evaluates MCQs effectiveness relative to open-response questions, both individually and in combination, on learning. These activities are embedded within six tutor lessons on advocacy. Using a posttest-only randomized control design, we compare the performance of 234 tutors (790 lesson completions) across three conditions: MCQ only, open response only, and a combination of both. We find no significant learning differences across conditions at posttest, but tutors in the MCQ condition took significantly less time to complete instruction. These findings suggest that MCQs are as effective, and more efficient, than open response tasks for learning when practice time is limited. To further enhance efficiency, we autograded open responses using GPT-4o and GPT-4-turbo. GPT models demonstrate proficiency for purposes of low-stakes assessment, though further research is needed for broader use. This study contributes a dataset of lesson log data, human annotation rubrics, and LLM prompts to promote transparency and reproducibility.
comment: Full research paper accepted to Learning Analytics and Knowledge (LAK 2025)
☆ Targeted Angular Reversal of Weights (TARS) for Knowledge Removal in Large Language Models
The sheer scale of data required to train modern large language models (LLMs) poses significant risks, as models are likely to gain knowledge of sensitive topics such as bio-security, as well the ability to replicate copyrighted works. Methods designed to remove such knowledge must do so from all prompt directions, in a multi-lingual capacity and without degrading general model performance. To this end, we introduce the targeted angular reversal (TARS) method of knowledge removal from LLMs. The TARS method firstly leverages the LLM in combination with a detailed prompt to aggregate information about a selected concept in the internal representation space of the LLM. It then refines this approximate concept vector to trigger the concept token with high probability, by perturbing the approximate concept vector with noise and transforming it into token scores with the language model head. The feedforward weight vectors in the LLM which operate directly on the internal representation space, and have the highest cosine similarity with this targeting vector, are then replaced by a reversed targeting vector, thus limiting the ability of the concept to propagate through the model. The modularity of the TARS method allows for a sequential removal of concepts from Llama 3.1 8B, such as the famous literary detective Sherlock Holmes, and the planet Saturn. It is demonstrated that the probability of triggering target concepts can be reduced to 0.00 with as few as 1 TARS edit, whilst simultaneously removing the knowledge bi-directionally. Moreover, knowledge is shown to be removed across all languages despite only being targeted in English. Importantly, TARS has minimal impact on the general model capabilities, as after removing 5 diverse concepts in a modular fashion, there is minimal KL divergence in the next token probabilities of the LLM on large corpora of Wikipedia text (median of 0.002).
comment: 14 pages, 5 figures, 1 table
☆ Exploring the Frontiers of Animation Video Generation in the Sora Era: Method, Dataset and Benchmark
Animation has gained significant interest in the recent film and TV industry. Despite the success of advanced video generation models like Sora, Kling, and CogVideoX in generating natural videos, they lack the same effectiveness in handling animation videos. Evaluating animation video generation is also a great challenge due to its unique artist styles, violating the laws of physics and exaggerated motions. In this paper, we present a comprehensive system, AniSora, designed for animation video generation, which includes a data processing pipeline, a controllable generation model, and an evaluation dataset. Supported by the data processing pipeline with over 10M high-quality data, the generation model incorporates a spatiotemporal mask module to facilitate key animation production functions such as image-to-video generation, frame interpolation, and localized image-guided animation. We also collect an evaluation benchmark of 948 various animation videos, the evaluation on VBench and human double-blind test demonstrates consistency in character and motion, achieving state-of-the-art results in animation video generation. %We also collect an evaluation benchmark of 948 various animation videos, with specifically developed metrics for animation video generation. Our model access API and evaluation benchmark will be publicly available.
☆ Physics Instrument Design with Reinforcement Learning
We present a case for the use of Reinforcement Learning (RL) for the design of physics instrument as an alternative to gradient-based instrument-optimization methods. It's applicability is demonstrated using two empirical studies. One is longitudinal segmentation of calorimeters and the second is both transverse segmentation as well longitudinal placement of trackers in a spectrometer. Based on these experiments, we propose an alternative approach that offers unique advantages over differentiable programming and surrogate-based differentiable design optimization methods. First, Reinforcement Learning (RL) algorithms possess inherent exploratory capabilities, which help mitigate the risk of convergence to local optima. Second, this approach eliminates the necessity of constraining the design to a predefined detector model with fixed parameters. Instead, it allows for the flexible placement of a variable number of detector components and facilitates discrete decision-making. We then discuss the road map of how this idea can be extended into designing very complex instruments. The presented study sets the stage for a novel framework in physics instrument design, offering a scalable and efficient framework that can be pivotal for future projects such as the Future Circular Collider (FCC), where most optimized detectors are essential for exploring physics at unprecedented energy scales.
☆ How good is my story? Towards quantitative metrics for evaluating LLM-generated XAI narratives
A rapidly developing application of LLMs in XAI is to convert quantitative explanations such as SHAP into user-friendly narratives to explain the decisions made by smaller prediction models. Evaluating the narratives without relying on human preference studies or surveys is becoming increasingly important in this field. In this work we propose a framework and explore several automated metrics to evaluate LLM-generated narratives for explanations of tabular classification tasks. We apply our approach to compare several state-of-the-art LLMs across different datasets and prompt types. As a demonstration of their utility, these metrics allow us to identify new challenges related to LLM hallucinations for XAI narratives.
☆ GAF: Gaussian Avatar Reconstruction from Monocular Videos via Multi-view Diffusion
We propose a novel approach for reconstructing animatable 3D Gaussian avatars from monocular videos captured by commodity devices like smartphones. Photorealistic 3D head avatar reconstruction from such recordings is challenging due to limited observations, which leaves unobserved regions under-constrained and can lead to artifacts in novel views. To address this problem, we introduce a multi-view head diffusion model, leveraging its priors to fill in missing regions and ensure view consistency in Gaussian splatting renderings. To enable precise viewpoint control, we use normal maps rendered from FLAME-based head reconstruction, which provides pixel-aligned inductive biases. We also condition the diffusion model on VAE features extracted from the input image to preserve details of facial identity and appearance. For Gaussian avatar reconstruction, we distill multi-view diffusion priors by using iteratively denoised images as pseudo-ground truths, effectively mitigating over-saturation issues. To further improve photorealism, we apply latent upsampling to refine the denoised latent before decoding it into an image. We evaluate our method on the NeRSemble dataset, showing that GAF outperforms the previous state-of-the-art methods in novel view synthesis by a 5.34\% higher SSIM score. Furthermore, we demonstrate higher-fidelity avatar reconstructions from monocular videos captured on commodity devices.
comment: Paper Video: https://youtu.be/QuIYTljvhyg Project Page: https://tangjiapeng.github.io/projects/GAF
☆ From Allies to Adversaries: Manipulating LLM Tool-Calling through Adversarial Injection
Tool-calling has changed Large Language Model (LLM) applications by integrating external tools, significantly enhancing their functionality across diverse tasks. However, this integration also introduces new security vulnerabilities, particularly in the tool scheduling mechanisms of LLM, which have not been extensively studied. To fill this gap, we present ToolCommander, a novel framework designed to exploit vulnerabilities in LLM tool-calling systems through adversarial tool injection. Our framework employs a well-designed two-stage attack strategy. Firstly, it injects malicious tools to collect user queries, then dynamically updates the injected tools based on the stolen information to enhance subsequent attacks. These stages enable ToolCommander to execute privacy theft, launch denial-of-service attacks, and even manipulate business competition by triggering unscheduled tool-calling. Notably, the ASR reaches 91.67% for privacy theft and hits 100% for denial-of-service and unscheduled tool calling in certain cases. Our work demonstrates that these vulnerabilities can lead to severe consequences beyond simple misuse of tool-calling systems, underscoring the urgent need for robust defensive strategies to secure LLM Tool-calling systems.
☆ BiCert: A Bilinear Mixed Integer Programming Formulation for Precise Certified Bounds Against Data Poisoning Attacks
Data poisoning attacks pose one of the biggest threats to modern AI systems, necessitating robust defenses. While extensive efforts have been made to develop empirical defenses, attackers continue to evolve, creating sophisticated methods to circumvent these measures. To address this, we must move beyond empirical defenses and establish provable certification methods that guarantee robustness. This paper introduces a novel certification approach, BiCert, using Bilinear Mixed Integer Programming (BMIP) to compute sound deterministic bounds that provide such provable robustness. Using BMIP, we compute the reachable set of parameters that could result from training with potentially manipulated data. A key element to make this computation feasible is to relax the reachable parameter set to a convex set between training iterations. At test time, this parameter set allows us to predict all possible outcomes, guaranteeing robustness. BiCert is more precise than previous methods, which rely solely on interval and polyhedral bounds. Crucially, our approach overcomes the fundamental limitation of prior approaches where parameter bounds could only grow, often uncontrollably. We show that BiCert's tighter bounds eliminate a key source of divergence issues, resulting in more stable training and higher certified accuracy.
☆ Solving Robust Markov Decision Processes: Generic, Reliable, Efficient AAAI'25
Markov decision processes (MDP) are a well-established model for sequential decision-making in the presence of probabilities. In robust MDP (RMDP), every action is associated with an uncertainty set of probability distributions, modelling that transition probabilities are not known precisely. Based on the known theoretical connection to stochastic games, we provide a framework for solving RMDPs that is generic, reliable, and efficient. It is *generic* both with respect to the model, allowing for a wide range of uncertainty sets, including but not limited to intervals, $L^1$- or $L^2$-balls, and polytopes; and with respect to the objective, including long-run average reward, undiscounted total reward, and stochastic shortest path. It is *reliable*, as our approach not only converges in the limit, but provides precision guarantees at any time during the computation. It is *efficient* because -- in contrast to state-of-the-art approaches -- it avoids explicitly constructing the underlying stochastic game. Consequently, our prototype implementation outperforms existing tools by several orders of magnitude and can solve RMDPs with a million states in under a minute.
comment: Accepted for publication at AAAI'25. Extended version with full appendix, 26 pages
☆ Multi-Head Encoding for Extreme Label Classification
The number of categories of instances in the real world is normally huge, and each instance may contain multiple labels. To distinguish these massive labels utilizing machine learning, eXtreme Label Classification (XLC) has been established. However, as the number of categories increases, the number of parameters and nonlinear operations in the classifier also rises. This results in a Classifier Computational Overload Problem (CCOP). To address this, we propose a Multi-Head Encoding (MHE) mechanism, which replaces the vanilla classifier with a multi-head classifier. During the training process, MHE decomposes extreme labels into the product of multiple short local labels, with each head trained on these local labels. During testing, the predicted labels can be directly calculated from the local predictions of each head. This reduces the computational load geometrically. Then, according to the characteristics of different XLC tasks, e.g., single-label, multi-label, and model pretraining tasks, three MHE-based implementations, i.e., Multi-Head Product, Multi-Head Cascade, and Multi-Head Sampling, are proposed to more effectively cope with CCOP. Moreover, we theoretically demonstrate that MHE can achieve performance approximately equivalent to that of the vanilla classifier by generalizing the low-rank approximation problem from Frobenius-norm to Cross-Entropy. Experimental results show that the proposed methods achieve state-of-the-art performance while significantly streamlining the training and inference processes of XLC tasks. The source code has been made public at https://github.com/Anoise/MHE.
comment: 20 pages, 12 figs, Published in TPAMI
☆ SwiftTry: Fast and Consistent Video Virtual Try-On with Diffusion Models
Given an input video of a person and a new garment, the objective of this paper is to synthesize a new video where the person is wearing the specified garment while maintaining spatiotemporal consistency. While significant advances have been made in image-based virtual try-ons, extending these successes to video often results in frame-to-frame inconsistencies. Some approaches have attempted to address this by increasing the overlap of frames across multiple video chunks, but this comes at a steep computational cost due to the repeated processing of the same frames, especially for long video sequence. To address these challenges, we reconceptualize video virtual try-on as a conditional video inpainting task, with garments serving as input conditions. Specifically, our approach enhances image diffusion models by incorporating temporal attention layers to improve temporal coherence. To reduce computational overhead, we introduce ShiftCaching, a novel technique that maintains temporal consistency while minimizing redundant computations. Furthermore, we introduce the \dataname~dataset, a new video try-on dataset featuring more complex backgrounds, challenging movements, and higher resolution compared to existing public datasets. Extensive experiments show that our approach outperforms current baselines, particularly in terms of video consistency and inference speed. Data and code are available at https://github.com/VinAIResearch/swift-try
☆ Scaling Combinatorial Optimization Neural Improvement Heuristics with Online Search and Adaptation
We introduce Limited Rollout Beam Search (LRBS), a beam search strategy for deep reinforcement learning (DRL) based combinatorial optimization improvement heuristics. Utilizing pre-trained models on the Euclidean Traveling Salesperson Problem, LRBS significantly enhances both in-distribution performance and generalization to larger problem instances, achieving optimality gaps that outperform existing improvement heuristics and narrowing the gap with state-of-the-art constructive methods. We also extend our analysis to two pickup and delivery TSP variants to validate our results. Finally, we employ our search strategy for offline and online adaptation of the pre-trained improvement policy, leading to improved search performance and surpassing recent adaptive methods for constructive heuristics.
☆ WordVIS: A Color Worth A Thousand Words
Document classification is considered a critical element in automated document processing systems. In recent years multi-modal approaches have become increasingly popular for document classification. Despite their improvements, these approaches are underutilized in the industry due to their requirement for a tremendous volume of training data and extensive computational power. In this paper, we attempt to address these issues by embedding textual features directly into the visual space, allowing lightweight image-based classifiers to achieve state-of-the-art results using small-scale datasets in document classification. To evaluate the efficacy of the visual features generated from our approach on limited data, we tested on the standard dataset Tobacco-3482. Our experiments show a tremendous improvement in image-based classifiers, achieving an improvement of 4.64% using ResNet50 with no document pre-training. It also sets a new record for the best accuracy of the Tobacco-3482 dataset with a score of 91.14% using the image-based DocXClassifier with no document pre-training. The simplicity of the approach, its resource requirements, and subsequent results provide a good prospect for its use in industrial use cases.
☆ Direct Encoding of Declare Constraints in ASP
Answer Set Programming (ASP), a well-known declarative logic programming paradigm, has recently found practical application in Process Mining. In particular, ASP has been used to model tasks involving declarative specifications of business processes. In this area, Declare stands out as the most widely adopted declarative process modeling language, offering a means to model processes through sets of constraints valid traces must satisfy, that can be expressed in Linear Temporal Logic over Finite Traces (LTLf). Existing ASP-based solutions encode Declare constraints by modeling the corresponding LTLf formula or its equivalent automaton which can be obtained using established techniques. In this paper, we introduce a novel encoding for Declare constraints that directly models their semantics as ASP rules, eliminating the need for intermediate representations. We assess the effectiveness of this novel approach on two Process Mining tasks by comparing it with alternative ASP encodings and a Python library for Declare. Under consideration in Theory and Practice of Logic Programming (TPLP).
comment: Under consideration in Theory and Practice of Logic Programming (TPLP)
☆ VLR-Bench: Multilingual Benchmark Dataset for Vision-Language Retrieval Augmented Generation COLING 2025
We propose the VLR-Bench, a visual question answering (VQA) benchmark for evaluating vision language models (VLMs) based on retrieval augmented generation (RAG). Unlike existing evaluation datasets for external knowledge-based VQA, the proposed VLR-Bench includes five input passages. This allows testing of the ability to determine which passage is useful for answering a given query, a capability lacking in previous research. In this context, we constructed a dataset of 32,000 automatically generated instruction-following examples, which we denote as VLR-IF. This dataset is specifically designed to enhance the RAG capabilities of VLMs by enabling them to learn how to generate appropriate answers based on input passages. We evaluated the validity of the proposed benchmark and training data and verified its performance using the state-of-the-art Llama3-based VLM, the Llava-Llama-3 model. The proposed VLR-Bench and VLR-IF datasets are publicly available online.
comment: The 31st International Conference on Computational Linguistics (COLING 2025), 19 pages
☆ ROUTE: Robust Multitask Tuning and Collaboration for Text-to-SQL
Despite the significant advancements in Text-to-SQL (Text2SQL) facilitated by large language models (LLMs), the latest state-of-the-art techniques are still trapped in the in-context learning of closed-source LLMs (e.g., GPT-4), which limits their applicability in open scenarios. To address this challenge, we propose a novel RObust mUltitask Tuning and collaboration mEthod (ROUTE) to improve the comprehensive capabilities of open-source LLMs for Text2SQL, thereby providing a more practical solution. Our approach begins with multi-task supervised fine-tuning (SFT) using various synthetic training data related to SQL generation. Unlike existing SFT-based Text2SQL methods, we introduced several additional SFT tasks, including schema linking, noise correction, and continuation writing. Engaging in a variety of SQL generation tasks enhances the model's understanding of SQL syntax and improves its ability to generate high-quality SQL queries. Additionally, inspired by the collaborative modes of LLM agents, we introduce a Multitask Collaboration Prompting (MCP) strategy. This strategy leverages collaboration across several SQL-related tasks to reduce hallucinations during SQL generation, thereby maximizing the potential of enhancing Text2SQL performance through explicit multitask capabilities. Extensive experiments and in-depth analyses have been performed on eight open-source LLMs and five widely-used benchmarks. The results demonstrate that our proposal outperforms the latest Text2SQL methods and yields leading performance.
☆ Can LLMs Convert Graphs to Text-Attributed Graphs?
Graphs are ubiquitous data structures found in numerous real-world applications, such as drug discovery, recommender systems, and social network analysis. Graph neural networks (GNNs) have become a popular tool to learn node embeddings through message passing on these structures. However, a significant challenge arises when applying GNNs to multiple graphs with different feature spaces, as existing GNN architectures are not designed for cross-graph feature alignment. To address this, recent approaches introduce text-attributed graphs, where each node is associated with a textual description, enabling the use of a shared textual encoder to project nodes from different graphs into a unified feature space. While promising, this method relies heavily on the availability of text-attributed data, which can be difficult to obtain in practice. To bridge this gap, we propose a novel method named Topology-Aware Node description Synthesis (TANS), which leverages large language models (LLMs) to automatically convert existing graphs into text-attributed graphs. The key idea is to integrate topological information with each node's properties, enhancing the LLMs' ability to explain how graph topology influences node semantics. We evaluate our TANS on text-rich, text-limited, and text-free graphs, demonstrating that it enables a single GNN to operate across diverse graphs. Notably, on text-free graphs, our method significantly outperforms existing approaches that manually design node features, showcasing the potential of LLMs for preprocessing graph-structured data, even in the absence of textual information. The code and data are available at https://github.com/Zehong-Wang/TANS.
☆ You Name It, I Run It: An LLM Agent to Execute Tests of Arbitrary Projects
The ability to execute the test suite of a project is essential in many scenarios, e.g., to assess code quality and code coverage, to validate code changes made by developers or automated tools, and to ensure compatibility with dependencies. Despite its importance, executing the test suite of a project can be challenging in practice because different projects use different programming languages, software ecosystems, build systems, testing frameworks, and other tools. These challenges make it difficult to create a reliable, universal test execution method that works across different projects. This paper presents ExecutionAgent, an automated technique that installs arbitrary projects, configures them to run test cases, and produces project-specific scripts to reproduce the setup. Inspired by the way a human developer would address this task, our approach is a large language model-based agent that autonomously executes commands and interacts with the host system. The agent uses meta-prompting to gather guidelines on the latest technologies related to the given project, and it iteratively refines its process based on feedback from the previous steps. Our evaluation applies ExecutionAgent to 50 open-source projects that use 14 different programming languages and many different build and testing tools. The approach successfully executes the test suites of 33/55 projects, while matching the test results of ground truth test suite executions with a deviation of only 7.5\%. These results improve over the best previously available technique by 6.6x. The costs imposed by the approach are reasonable, with an execution time of 74 minutes and LLM costs of 0.16 dollars, on average per project. We envision ExecutionAgent to serve as a valuable tool for developers, automated programming tools, and researchers that need to execute tests across a wide variety of projects.
☆ CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Models
In our previous work, we introduced CosyVoice, a multilingual speech synthesis model based on supervised discrete speech tokens. By employing progressive semantic decoding with two popular generative models, language models (LMs) and Flow Matching, CosyVoice demonstrated high prosody naturalness, content consistency, and speaker similarity in speech in-context learning. Recently, significant progress has been made in multi-modal large language models (LLMs), where the response latency and real-time factor of speech synthesis play a crucial role in the interactive experience. Therefore, in this report, we present an improved streaming speech synthesis model, CosyVoice 2, which incorporates comprehensive and systematic optimizations. Specifically, we introduce finite-scalar quantization to improve the codebook utilization of speech tokens. For the text-speech LM, we streamline the model architecture to allow direct use of a pre-trained LLM as the backbone. In addition, we develop a chunk-aware causal flow matching model to support various synthesis scenarios, enabling both streaming and non-streaming synthesis within a single model. By training on a large-scale multilingual dataset, CosyVoice 2 achieves human-parity naturalness, minimal response latency, and virtually lossless synthesis quality in the streaming mode. We invite readers to listen to the demos at https://funaudiollm.github.io/cosyvoice2.
comment: Tech report, work in progress
☆ Label-template based Few-Shot Text Classification with Contrastive Learning
As an algorithmic framework for learning to learn, meta-learning provides a promising solution for few-shot text classification. However, most existing research fail to give enough attention to class labels. Traditional basic framework building meta-learner based on prototype networks heavily relies on inter-class variance, and it is easily influenced by noise. To address these limitations, we proposes a simple and effective few-shot text classification framework. In particular, the corresponding label templates are embed into input sentences to fully utilize the potential value of class labels, guiding the pre-trained model to generate more discriminative text representations through the semantic information conveyed by labels. With the continuous influence of label semantics, supervised contrastive learning is utilized to model the interaction information between support samples and query samples. Furthermore, the averaging mechanism is replaced with an attention mechanism to highlight vital semantic information. To verify the proposed scheme, four typical datasets are employed to assess the performance of different methods. Experimental results demonstrate that our method achieves substantial performance enhancements and outperforms existing state-of-the-art models on few-shot text classification tasks.
☆ NetOrchLLM: Mastering Wireless Network Orchestration with Large Language Models
The transition to 6G networks promises unprecedented advancements in wireless communication, with increased data rates, ultra-low latency, and enhanced capacity. However, the complexity of managing and optimizing these next-generation networks presents significant challenges. The advent of large language models (LLMs) has revolutionized various domains by leveraging their sophisticated natural language understanding capabilities. However, the practical application of LLMs in wireless network orchestration and management remains largely unexplored. Existing literature predominantly offers visionary perspectives without concrete implementations, leaving a significant gap in the field. To address this gap, this paper presents NETORCHLLM, a wireless NETwork ORCHestrator LLM framework that uses LLMs to seamlessly orchestrate diverse wireless-specific models from wireless communication communities using their language understanding and generation capabilities. A comprehensive framework is introduced, demonstrating the practical viability of our approach and showcasing how LLMs can be effectively harnessed to optimize dense network operations, manage dynamic environments, and improve overall network performance. NETORCHLLM bridges the theoretical aspirations of prior research with practical, actionable solutions, paving the way for future advancements in integrating generative AI technologies within the wireless communications sector.
☆ A Cascaded Dilated Convolution Approach for Mpox Lesion Classification
The global outbreak of Mpox virus, classified as a Public Health Emergency of International Concern by WHO, presents significant diagnostic challenges due to its visual similarity to other skin lesion diseases. Current clinical detection techniques face limitations in accuracy and efficiency, necessitating improved automated diagnostic solutions. This study introduces a novel Cascaded Atrous Group Attention (CAGA) module, specifically designed to enhance multi-scale feature representation while optimizing computational efficiency. By integrating CAGA with EfficientViT-L1 as the backbone architecture, our approach achieves state-of-the-art performance with a score of 0.98% on the MCSI dataset, while reducing model parameters by 37.5% compared to the original EfficientViT-L1. This reduction in computational complexity maintains diagnostic accuracy while enabling broader deployment across resource-constrained healthcare settings. Extensive validation across two other benchmark datasets, including MSID and MSLD, demonstrate the model's robustness, consistently outperforming existing approaches. Our findings suggest that CAGA's efficient feature extraction mechanism could be adapted for other medical imaging tasks requiring fine-grained visual discrimination.
comment: (7 pages, 2 figures, 5 tables)
☆ RETQA: A Large-Scale Open-Domain Tabular Question Answering Dataset for Real Estate Sector AAAI 2025
The real estate market relies heavily on structured data, such as property details, market trends, and price fluctuations. However, the lack of specialized Tabular Question Answering datasets in this domain limits the development of automated question-answering systems. To fill this gap, we introduce RETQA, the first large-scale open-domain Chinese Tabular Question Answering dataset for Real Estate. RETQA comprises 4,932 tables and 20,762 question-answer pairs across 16 sub-fields within three major domains: property information, real estate company finance information and land auction information. Compared with existing tabular question answering datasets, RETQA poses greater challenges due to three key factors: long-table structures, open-domain retrieval, and multi-domain queries. To tackle these challenges, we propose the SLUTQA framework, which integrates large language models with spoken language understanding tasks to enhance retrieval and answering accuracy. Extensive experiments demonstrate that SLUTQA significantly improves the performance of large language models on RETQA by in-context learning. RETQA and SLUTQA provide essential resources for advancing tabular question answering research in the real estate domain, addressing critical challenges in open-domain and long-table question-answering. The dataset and code are publicly available at \url{https://github.com/jensen-w/RETQA}.
comment: This paper is accepted by AAAI 2025
☆ HiTZ at VarDial 2025 NorSID: Overcoming Data Scarcity with Language Transfer and Automatic Data Annotation
In this paper we present our submission for the NorSID Shared Task as part of the 2025 VarDial Workshop (Scherrer et al., 2025), consisting of three tasks: Intent Detection, Slot Filling and Dialect Identification, evaluated using data in different dialects of the Norwegian language. For Intent Detection and Slot Filling, we have fine-tuned a multitask model in a cross-lingual setting, to leverage the xSID dataset available in 17 languages. In the case of Dialect Identification, our final submission consists of a model fine-tuned on the provided development set, which has obtained the highest scores within our experiments. Our final results on the test set show that our models do not drop in performance compared to the development set, likely due to the domain-specificity of the dataset and the similar distribution of both subsets. Finally, we also report an in-depth analysis of the provided datasets and their artifacts, as well as other sets of experiments that have been carried out but did not yield the best results. Additionally, we present an analysis on the reasons why some methods have been more successful than others; mainly the impact of the combination of languages and domain-specificity of the training data on the results.
comment: Vardial 2025 NorSID Shared Task
☆ AI in the Cosmos
Artificial intelligence (AI) is revolutionizing research by enabling the efficient analysis of large datasets and the discovery of hidden patterns. In astrophysics, AI has become essential, transforming the classification of celestial sources, data modeling, and the interpretation of observations. In this review, I highlight examples of AI applications in astrophysics, including source classification, spectral energy distribution modeling, and discuss the advancements achievable through generative AI. However, the use of AI introduces challenges, including biases, errors, and the "black box" nature of AI models, which must be resolved before their application. These issues can be addressed through the concept of Human-Guided AI (HG-AI), which integrates human expertise and domain-specific knowledge into AI applications. This approach aims to ensure that AI is applied in a robust, interpretable, and ethical manner, leading to deeper insights and fostering scientific excellence.
comment: In press in the International Journal of Modern Physics D; invited talk at the 17th Marcel Grossmann Meeting
☆ Data Pruning Can Do More: A Comprehensive Data Pruning Approach for Object Re-identification
Previous studies have demonstrated that not each sample in a dataset is of equal importance during training. Data pruning aims to remove less important or informative samples while still achieving comparable results as training on the original (untruncated) dataset, thereby reducing storage and training costs. However, the majority of data pruning methods are applied to image classification tasks. To our knowledge, this work is the first to explore the feasibility of these pruning methods applied to object re-identification (ReID) tasks, while also presenting a more comprehensive data pruning approach. By fully leveraging the logit history during training, our approach offers a more accurate and comprehensive metric for quantifying sample importance, as well as correcting mislabeled samples and recognizing outliers. Furthermore, our approach is highly efficient, reducing the cost of importance score estimation by 10 times compared to existing methods. Our approach is a plug-and-play, architecture-agnostic framework that can eliminate/reduce 35%, 30%, and 5% of samples/training time on the VeRi, MSMT17 and Market1501 datasets, respectively, with negligible loss in accuracy (< 0.1%). The lists of important, mislabeled, and outlier samples from these ReID datasets are available at https://github.com/Zi-Y/data-pruning-reid.
☆ Panacea: Novel DNN Accelerator using Accuracy-Preserving Asymmetric Quantization and Energy-Saving Bit-Slice Sparsity HPCA 2025
Low bit-precisions and their bit-slice sparsity have recently been studied to accelerate general matrix-multiplications (GEMM) during large-scale deep neural network (DNN) inferences. While the conventional symmetric quantization facilitates low-resolution processing with bit-slice sparsity for both weight and activation, its accuracy loss caused by the activation's asymmetric distributions cannot be acceptable, especially for large-scale DNNs. In efforts to mitigate this accuracy loss, recent studies have actively utilized asymmetric quantization for activations without requiring additional operations. However, the cutting-edge asymmetric quantization produces numerous nonzero slices that cannot be compressed and skipped by recent bit-slice GEMM accelerators, naturally consuming more processing energy to handle the quantized DNN models. To simultaneously achieve high accuracy and hardware efficiency for large-scale DNN inferences, this paper proposes an Asymmetrically-Quantized bit-Slice GEMM (AQS-GEMM) for the first time. In contrast to the previous bit-slice computing, which only skips operations of zero slices, the AQS-GEMM compresses frequent nonzero slices, generated by asymmetric quantization, and skips their operations. To increase the slice-level sparsity of activations, we also introduce two algorithm-hardware co-optimization methods: a zero-point manipulation and a distribution-based bit-slicing. To support the proposed AQS-GEMM and optimizations at the hardware-level, we newly introduce a DNN accelerator, Panacea, which efficiently handles sparse/dense workloads of the tiled AQS-GEMM to increase data reuse and utilization. Panacea supports a specialized dataflow and run-length encoding to maximize data reuse and minimize external memory accesses, significantly improving its hardware efficiency. Our benchmark evaluations show Panacea outperforms existing DNN accelerators.
comment: 15 pages, 20 figures, Accepted to HPCA 2025
☆ GAOKAO-Eval: Does high scores truly reflect strong capabilities in LLMs?
Large Language Models (LLMs) are commonly evaluated using human-crafted benchmarks, under the premise that higher scores implicitly reflect stronger human-like performance. However, there is growing concern that LLMs may ``game" these benchmarks due to data leakage, achieving high scores while struggling with tasks simple for humans. To substantively address the problem, we create GAOKAO-Eval, a comprehensive benchmark based on China's National College Entrance Examination (Gaokao), and conduct ``closed-book" evaluations for representative models released prior to Gaokao. Contrary to prevailing consensus, even after addressing data leakage and comprehensiveness, GAOKAO-Eval reveals that high scores still fail to truly reflect human-aligned capabilities. To better understand this mismatch, We introduce the Rasch model from cognitive psychology to analyze LLM scoring patterns and identify two key discrepancies: 1) anomalous consistent performance across various question difficulties, and 2) high variance in performance on questions of similar difficulty. In addition, We identified inconsistent grading of LLM-generated answers among teachers and recurring mistake patterns. we find that the phenomenons are well-grounded in the motivations behind OpenAI o1, and o1's reasoning-as-difficulties can mitigate the mismatch. These results show that GAOKAO-Eval can reveal limitations in LLM capabilities not captured by current benchmarks and highlight the need for more LLM-aligned difficulty analysis.
comment: 10 pages, 13 figures
☆ TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
Recent advances in Gaussian Splatting have significantly advanced the field, achieving both panoptic and interactive segmentation of 3D scenes. However, existing methodologies often overlook the critical need for reconstructing specified targets with complex structures from sparse views. To address this issue, we introduce TSGaussian, a novel framework that combines semantic constraints with depth priors to avoid geometry degradation in challenging novel view synthesis tasks. Our approach prioritizes computational resources on designated targets while minimizing background allocation. Bounding boxes from YOLOv9 serve as prompts for Segment Anything Model to generate 2D mask predictions, ensuring semantic accuracy and cost efficiency. TSGaussian effectively clusters 3D gaussians by introducing a compact identity encoding for each Gaussian ellipsoid and incorporating 3D spatial consistency regularization. Leveraging these modules, we propose a pruning strategy to effectively reduce redundancy in 3D gaussians. Extensive experiments demonstrate that TSGaussian outperforms state-of-the-art methods on three standard datasets and a new challenging dataset we collected, achieving superior results in novel view synthesis of specific objects. Code is available at: https://github.com/leon2000-ai/TSGaussian.
☆ Large Action Models: From Inception to Implementation
As AI continues to advance, there is a growing demand for systems that go beyond language-based assistance and move toward intelligent agents capable of performing real-world actions. This evolution requires the transition from traditional Large Language Models (LLMs), which excel at generating textual responses, to Large Action Models (LAMs), designed for action generation and execution within dynamic environments. Enabled by agent systems, LAMs hold the potential to transform AI from passive language understanding to active task completion, marking a significant milestone in the progression toward artificial general intelligence. In this paper, we present a comprehensive framework for developing LAMs, offering a systematic approach to their creation, from inception to deployment. We begin with an overview of LAMs, highlighting their unique characteristics and delineating their differences from LLMs. Using a Windows OS-based agent as a case study, we provide a detailed, step-by-step guide on the key stages of LAM development, including data collection, model training, environment integration, grounding, and evaluation. This generalizable workflow can serve as a blueprint for creating functional LAMs in various application domains. We conclude by identifying the current limitations of LAMs and discussing directions for future research and industrial deployment, emphasizing the challenges and opportunities that lie ahead in realizing the full potential of LAMs in real-world applications. The code for the data collection process utilized in this paper is publicly available at: https://github.com/microsoft/UFO/tree/main/dataflow, and comprehensive documentation can be found at https://microsoft.github.io/UFO/dataflow/overview/.
comment: 25pages,12 figures
☆ Enhanced Speech Emotion Recognition with Efficient Channel Attention Guided Deep CNN-BiLSTM Framework
Speech emotion recognition (SER) is crucial for enhancing affective computing and enriching the domain of human-computer interaction. However, the main challenge in SER lies in selecting relevant feature representations from speech signals with lower computational costs. In this paper, we propose a lightweight SER architecture that integrates attention-based local feature blocks (ALFBs) to capture high-level relevant feature vectors from speech signals. We also incorporate a global feature block (GFB) technique to capture sequential, global information and long-term dependencies in speech signals. By aggregating attention-based local and global contextual feature vectors, our model effectively captures the internal correlation between salient features that reflect complex human emotional cues. To evaluate our approach, we extracted four types of spectral features from speech audio samples: mel-frequency cepstral coefficients, mel-spectrogram, root mean square value, and zero-crossing rate. Through a 5-fold cross-validation strategy, we tested the proposed method on five multi-lingual standard benchmark datasets: TESS, RAVDESS, BanglaSER, SUBESCO, and Emo-DB, and obtained a mean accuracy of 99.65%, 94.88%, 98.12%, 97.94%, and 97.19% respectively. The results indicate that our model achieves state-of-the-art (SOTA) performance compared to most existing methods.
comment: 42 pages,10 figures
☆ Cycle-Consistent Bridge Diffusion Model for Accelerated MRI Reconstruction
Accelerated MRI reconstruction techniques aim to reduce examination time while maintaining high image fidelity, which is highly desirable in clinical settings for improving patient comfort and hospital efficiency. Existing deep learning methods typically reconstruct images from under-sampled data with traditional reconstruction approaches, but they still struggle to provide high-fidelity results. Diffusion models show great potential to improve fidelity of generated images in recent years. However, their inference process starting with a random Gaussian noise introduces instability into the results and usually requires thousands of sampling steps, resulting in sub-optimal reconstruction quality and low efficiency. To address these challenges, we propose Cycle-Consistent Bridge Diffusion Model (CBDM). CBDM employs two bridge diffusion models to construct a cycle-consistent diffusion process with a consistency loss, enhancing the fine-grained details of reconstructed images and reducing the number of diffusion steps. Moreover, CBDM incorporates a Contourlet Decomposition Embedding Module (CDEM) which captures multi-scale structural texture knowledge in images through frequency domain decomposition pyramids and directional filter banks to improve structural fidelity. Extensive experiments demonstrate the superiority of our model by higher reconstruction quality and fewer training iterations, achieving a new state of the art for accelerated MRI reconstruction in both fastMRI and IXI datasets.
☆ Visual Object Tracking across Diverse Data Modalities: A Review
Visual Object Tracking (VOT) is an attractive and significant research area in computer vision, which aims to recognize and track specific targets in video sequences where the target objects are arbitrary and class-agnostic. The VOT technology could be applied in various scenarios, processing data of diverse modalities such as RGB, thermal infrared and point cloud. Besides, since no one sensor could handle all the dynamic and varying environments, multi-modal VOT is also investigated. This paper presents a comprehensive survey of the recent progress of both single-modal and multi-modal VOT, especially the deep learning methods. Specifically, we first review three types of mainstream single-modal VOT, including RGB, thermal infrared and point cloud tracking. In particular, we conclude four widely-used single-modal frameworks, abstracting their schemas and categorizing the existing inheritors. Then we summarize four kinds of multi-modal VOT, including RGB-Depth, RGB-Thermal, RGB-LiDAR and RGB-Language. Moreover, the comparison results in plenty of VOT benchmarks of the discussed modalities are presented. Finally, we provide recommendations and insightful observations, inspiring the future development of this fast-growing literature.
☆ Small Language Model as Data Prospector for Large Language Model
The quality of instruction data directly affects the performance of fine-tuned Large Language Models (LLMs). Previously, \cite{li2023one} proposed \texttt{NUGGETS}, which identifies and selects high-quality quality data from a large dataset by identifying those individual instruction examples that can significantly improve the performance of different tasks after being learnt as one-shot instances. In this work, we propose \texttt{SuperNUGGETS}, an improved variant of \texttt{NUGGETS} optimised for efficiency and performance. Our \texttt{SuperNUGGETS} uses a small language model (SLM) instead of a large language model (LLM) to filter the data for outstanding one-shot instances and refines the predefined set of tests. The experimental results show that the performance of \texttt{SuperNUGGETS} only decreases by 1-2% compared to \texttt{NUGGETS}, but the efficiency can be increased by a factor of 58. Compared to the original \texttt{NUGGETS}, our \texttt{SuperNUGGETS} has a higher utility value due to the significantly lower resource consumption.
☆ One Filter to Deploy Them All: Robust Safety for Quadrupedal Navigation in Unknown Environments
As learning-based methods for legged robots rapidly grow in popularity, it is important that we can provide safety assurances efficiently across different controllers and environments. Existing works either rely on a priori knowledge of the environment and safety constraints to ensure system safety or provide assurances for a specific locomotion policy. To address these limitations, we propose an observation-conditioned reachability-based (OCR) safety-filter framework. Our key idea is to use an OCR value network (OCR-VN) that predicts the optimal control-theoretic safety value function for new failure regions and dynamic uncertainty during deployment time. Specifically, the OCR-VN facilitates rapid safety adaptation through two key components: a LiDAR-based input that allows the dynamic construction of safe regions in light of new obstacles and a disturbance estimation module that accounts for dynamics uncertainty in the wild. The predicted safety value function is used to construct an adaptive safety filter that overrides the nominal quadruped controller when necessary to maintain safety. Through simulation studies and hardware experiments on a Unitree Go1 quadruped, we demonstrate that the proposed framework can automatically safeguard a wide range of hierarchical quadruped controllers, adapts to novel environments, and is robust to unmodeled dynamics without a priori access to the controllers or environments - hence, "One Filter to Deploy Them All". The experiment videos can be found on the project website.
comment: Project website: https://sia-lab-git.github.io/One_Filter_to_Deploy_Them_All/
☆ AI and the Future of Digital Public Squares
Two substantial technological advances have reshaped the public square in recent decades: first with the advent of the internet and second with the recent introduction of large language models (LLMs). LLMs offer opportunities for a paradigm shift towards more decentralized, participatory online spaces that can be used to facilitate deliberative dialogues at scale, but also create risks of exacerbating societal schisms. Here, we explore four applications of LLMs to improve digital public squares: collective dialogue systems, bridging systems, community moderation, and proof-of-humanity systems. Building on the input from over 70 civil society experts and technologists, we argue that LLMs both afford promising opportunities to shift the paradigm for conversations at scale and pose distinct risks for digital public squares. We lay out an agenda for future research and investments in AI that will strengthen digital public squares and safeguard against potential misuses of AI.
comment: 40 pages, 5 figures
☆ SUMI-IFL: An Information-Theoretic Framework for Image Forgery Localization with Sufficiency and Minimality Constraints
Image forgery localization (IFL) is a crucial technique for preventing tampered image misuse and protecting social safety. However, due to the rapid development of image tampering technologies, extracting more comprehensive and accurate forgery clues remains an urgent challenge. To address these challenges, we introduce a novel information-theoretic IFL framework named SUMI-IFL that imposes sufficiency-view and minimality-view constraints on forgery feature representation. First, grounded in the theoretical analysis of mutual information, the sufficiency-view constraint is enforced on the feature extraction network to ensure that the latent forgery feature contains comprehensive forgery clues. Considering that forgery clues obtained from a single aspect alone may be incomplete, we construct the latent forgery feature by integrating several individual forgery features from multiple perspectives. Second, based on the information bottleneck, the minimality-view constraint is imposed on the feature reasoning network to achieve an accurate and concise forgery feature representation that counters the interference of task-unrelated features. Extensive experiments show the superior performance of SUMI-IFL to existing state-of-the-art methods, not only on in-dataset comparisons but also on cross-dataset comparisons.
☆ Efficient Large-Scale Traffic Forecasting with Transformers: A Spatial Data Management Perspective KDD 2025
Road traffic forecasting is crucial in real-world intelligent transportation scenarios like traffic dispatching and path planning in city management and personal traveling. Spatio-temporal graph neural networks (STGNNs) stand out as the mainstream solution in this task. Nevertheless, the quadratic complexity of remarkable dynamic spatial modeling-based STGNNs has become the bottleneck over large-scale traffic data. From the spatial data management perspective, we present a novel Transformer framework called PatchSTG to efficiently and dynamically model spatial dependencies for large-scale traffic forecasting with interpretability and fidelity. Specifically, we design a novel irregular spatial patching to reduce the number of points involved in the dynamic calculation of Transformer. The irregular spatial patching first utilizes the leaf K-dimensional tree (KDTree) to recursively partition irregularly distributed traffic points into leaf nodes with a small capacity, and then merges leaf nodes belonging to the same subtree into occupancy-equaled and non-overlapped patches through padding and backtracking. Based on the patched data, depth and breadth attention are used interchangeably in the encoder to dynamically learn local and global spatial knowledge from points in a patch and points with the same index of patches. Experimental results on four real world large-scale traffic datasets show that our PatchSTG achieves train speed and memory utilization improvements up to $10\times$ and $4\times$ with the state-of-the-art performance.
comment: Accepted by SIGKDD 2025
☆ EP-CFG: Energy-Preserving Classifier-Free Guidance
Classifier-free guidance (CFG) is widely used in diffusion models but often introduces over-contrast and over-saturation artifacts at higher guidance strengths. We present EP-CFG (Energy-Preserving Classifier-Free Guidance), which addresses these issues by preserving the energy distribution of the conditional prediction during the guidance process. Our method simply rescales the energy of the guided output to match that of the conditional prediction at each denoising step, with an optional robust variant for improved artifact suppression. Through experiments, we show that EP-CFG maintains natural image quality and preserves details across guidance strengths while retaining CFG's semantic alignment benefits, all with minimal computational overhead.
☆ What constitutes a Deep Fake? The blurry line between legitimate processing and manipulation under the EU AI Act
When does a digital image resemble reality? The relevance of this question increases as the generation of synthetic images -- so called deep fakes -- becomes increasingly popular. Deep fakes have gained much attention for a number of reasons -- among others, due to their potential to disrupt the political climate. In order to mitigate these threats, the EU AI Act implements specific transparency regulations for generating synthetic content or manipulating existing content. However, the distinction between real and synthetic images is -- even from a computer vision perspective -- far from trivial. We argue that the current definition of deep fakes in the AI act and the corresponding obligations are not sufficiently specified to tackle the challenges posed by deep fakes. By analyzing the life cycle of a digital photo from the camera sensor to the digital editing features, we find that: (1.) Deep fakes are ill-defined in the EU AI Act. The definition leaves too much scope for what a deep fake is. (2.) It is unclear how editing functions like Google's ``best take'' feature can be considered as an exception to transparency obligations. (3.) The exception for substantially edited images raises questions about what constitutes substantial editing of content and whether or not this editing must be perceptible by a natural person. Our results demonstrate that complying with the current AI Act transparency obligations is difficult for providers and deployers. As a consequence of the unclear provisions, there is a risk that exceptions may be either too broad or too limited. We intend our analysis to foster the discussion on what constitutes a deep fake and to raise awareness about the pitfalls in the current AI Act transparency obligations.
comment: Preprint. Accepted at ACM CS&Law '25
☆ Enhancing Nursing and Elderly Care with Large Language Models: An AI-Driven Framework
This paper explores the application of large language models (LLMs) in nursing and elderly care, focusing on AI-driven patient monitoring and interaction. We introduce a novel Chinese nursing dataset and implement incremental pre-training (IPT) and supervised fine-tuning (SFT) techniques to enhance LLM performance in specialized tasks. Using LangChain, we develop a dynamic nursing assistant capable of real-time care and personalized interventions. Experimental results demonstrate significant improvements, paving the way for AI-driven solutions to meet the growing demands of healthcare in aging populations.
☆ B-VLLM: A Vision Large Language Model with Balanced Spatio-Temporal Tokens
Recently, Vision Large Language Models (VLLMs) integrated with vision encoders have shown promising performance in vision understanding. The key of VLLMs is to encode visual content into sequences of visual tokens, enabling VLLMs to simultaneously process both visual and textual content. However, understanding videos, especially long videos, remain a challenge to VLLMs as the number of visual tokens grows rapidly when encoding videos, resulting in the risk of exceeding the context window of VLLMs and introducing heavy computation burden. To restrict the number of visual tokens, existing VLLMs either: (1) uniformly downsample videos into a fixed number of frames or (2) reducing the number of visual tokens encoded from each frame. We argue the former solution neglects the rich temporal cue in videos and the later overlooks the spatial details in each frame. In this work, we present Balanced-VLLM (B-VLLM): a novel VLLM framework that aims to effectively leverage task relevant spatio-temporal cues while restricting the number of visual tokens under the VLLM context window length. At the core of our method, we devise a text-conditioned adaptive frame selection module to identify frames relevant to the visual understanding task. The selected frames are then de-duplicated using a temporal frame token merging technique. The visual tokens of the selected frames are processed through a spatial token sampling module and an optional spatial token merging strategy to achieve precise control over the token count. Experimental results show that B-VLLM is effective in balancing the number of frames and visual tokens in video understanding, yielding superior performance on various video understanding benchmarks. Our code is available at https://github.com/zhuqiangLu/B-VLLM.
☆ Analyzing Fairness of Classification Machine Learning Model with Structured Dataset
Machine learning (ML) algorithms have become integral to decision making in various domains, including healthcare, finance, education, and law enforcement. However, concerns about fairness and bias in these systems pose significant ethical and social challenges. This study investigates the fairness of ML models applied to structured datasets in classification tasks, highlighting the potential for biased predictions to perpetuate systemic inequalities. A publicly available dataset from Kaggle was selected for analysis, offering a realistic scenario for evaluating fairness in machine learning workflows. To assess and mitigate biases, three prominent fairness libraries; Fairlearn by Microsoft, AIF360 by IBM, and the What If Tool by Google were employed. These libraries provide robust frameworks for analyzing fairness, offering tools to evaluate metrics, visualize results, and implement bias mitigation strategies. The research aims to assess the extent of bias in the ML models, compare the effectiveness of these libraries, and derive actionable insights for practitioners. The findings reveal that each library has unique strengths and limitations in fairness evaluation and mitigation. By systematically comparing their capabilities, this study contributes to the growing field of ML fairness by providing practical guidance for integrating fairness tools into real world applications. These insights are intended to support the development of more equitable machine learning systems.
comment: 12 pages, 3 tables
☆ Semi-Periodic Activation for Time Series Classification
This paper investigates the lack of research on activation functions for neural network models in time series tasks. It highlights the need to identify essential properties of these activations to improve their effectiveness in specific domains. To this end, the study comprehensively analyzes properties, such as bounded, monotonic, nonlinearity, and periodicity, for activation in time series neural networks. We propose a new activation that maximizes the coverage of these properties, called LeakySineLU. We empirically evaluate the LeakySineLU against commonly used activations in the literature using 112 benchmark datasets for time series classification, obtaining the best average ranking in all comparative scenarios.
☆ CSL-L2M: Controllable Song-Level Lyric-to-Melody Generation Based on Conditional Transformer with Fine-Grained Lyric and Musical Controls AAAI-25
Lyric-to-melody generation is a highly challenging task in the field of AI music generation. Due to the difficulty of learning strict yet weak correlations between lyrics and melodies, previous methods have suffered from weak controllability, low-quality and poorly structured generation. To address these challenges, we propose CSL-L2M, a controllable song-level lyric-to-melody generation method based on an in-attention Transformer decoder with fine-grained lyric and musical controls, which is able to generate full-song melodies matched with the given lyrics and user-specified musical attributes. Specifically, we first introduce REMI-Aligned, a novel music representation that incorporates strict syllable- and sentence-level alignments between lyrics and melodies, facilitating precise alignment modeling. Subsequently, sentence-level semantic lyric embeddings independently extracted from a sentence-wise Transformer encoder are combined with word-level part-of-speech embeddings and syllable-level tone embeddings as fine-grained controls to enhance the controllability of lyrics over melody generation. Then we introduce human-labeled musical tags, sentence-level statistical musical attributes, and learned musical features extracted from a pre-trained VQ-VAE as coarse-grained, fine-grained and high-fidelity controls, respectively, to the generation process, thereby enabling user control over melody generation. Finally, an in-attention Transformer decoder technique is leveraged to exert fine-grained control over the full-song melody generation with the aforementioned lyric and musical conditions. Experimental results demonstrate that our proposed CSL-L2M outperforms the state-of-the-art models, generating melodies with higher quality, better controllability and enhanced structure. Demos and source code are available at https://lichaiustc.github.io/CSL-L2M/.
comment: Accepted at AAAI-25
☆ Brain-inspired Chaotic Graph Backpropagation for Large-scale Combinatorial Optimization
Graph neural networks (GNNs) with unsupervised learning can solve large-scale combinatorial optimization problems (COPs) with efficient time complexity, making them versatile for various applications. However, since this method maps the combinatorial optimization problem to the training process of a graph neural network, and the current mainstream backpropagation-based training algorithms are prone to fall into local minima, the optimization performance is still inferior to the current state-of-the-art (SOTA) COP methods. To address this issue, inspired by possibly chaotic dynamics of real brain learning, we introduce a chaotic training algorithm, i.e. chaotic graph backpropagation (CGBP), which introduces a local loss function in GNN that makes the training process not only chaotic but also highly efficient. Different from existing methods, we show that the global ergodicity and pseudo-randomness of such chaotic dynamics enable CGBP to learn each optimal GNN effectively and globally, thus solving the COP efficiently. We have applied CGBP to solve various COPs, such as the maximum independent set, maximum cut, and graph coloring. Results on several large-scale benchmark datasets showcase that CGBP can outperform not only existing GNN algorithms but also SOTA methods. In addition to solving large-scale COPs, CGBP as a universal learning algorithm for GNNs, i.e. as a plug-in unit, can be easily integrated into any existing method for improving the performance.
☆ RLDG: Robotic Generalist Policy Distillation via Reinforcement Learning
Recent advances in robotic foundation models have enabled the development of generalist policies that can adapt to diverse tasks. While these models show impressive flexibility, their performance heavily depends on the quality of their training data. In this work, we propose Reinforcement Learning Distilled Generalists (RLDG), a method that leverages reinforcement learning to generate high-quality training data for finetuning generalist policies. Through extensive real-world experiments on precise manipulation tasks like connector insertion and assembly, we demonstrate that generalist policies trained with RL-generated data consistently outperform those trained with human demonstrations, achieving up to 40% higher success rates while generalizing better to new tasks. We also provide a detailed analysis that reveals this performance gain stems from both optimized action distributions and improved state coverage. Our results suggest that combining task-specific RL with generalist policy distillation offers a promising approach for developing more capable and efficient robotic manipulation systems that maintain the flexibility of foundation models while achieving the performance of specialized controllers. Videos and code can be found on our project website https://generalist-distillation.github.io
☆ LinGen: Towards High-Resolution Minute-Length Text-to-Video Generation with Linear Computational Complexity
Text-to-video generation enhances content creation but is highly computationally intensive: The computational cost of Diffusion Transformers (DiTs) scales quadratically in the number of pixels. This makes minute-length video generation extremely expensive, limiting most existing models to generating videos of only 10-20 seconds length. We propose a Linear-complexity text-to-video Generation (LinGen) framework whose cost scales linearly in the number of pixels. For the first time, LinGen enables high-resolution minute-length video generation on a single GPU without compromising quality. It replaces the computationally-dominant and quadratic-complexity block, self-attention, with a linear-complexity block called MATE, which consists of an MA-branch and a TE-branch. The MA-branch targets short-to-long-range correlations, combining a bidirectional Mamba2 block with our token rearrangement method, Rotary Major Scan, and our review tokens developed for long video generation. The TE-branch is a novel TEmporal Swin Attention block that focuses on temporal correlations between adjacent tokens and medium-range tokens. The MATE block addresses the adjacency preservation issue of Mamba and improves the consistency of generated videos significantly. Experimental results show that LinGen outperforms DiT (with a 75.6% win rate) in video quality with up to 15$\times$ (11.5$\times$) FLOPs (latency) reduction. Furthermore, both automatic metrics and human evaluation demonstrate our LinGen-4B yields comparable video quality to state-of-the-art models (with a 50.5%, 52.1%, 49.1% win rate with respect to Gen-3, LumaLabs, and Kling, respectively). This paves the way to hour-length movie generation and real-time interactive video generation. We provide 68s video generation results and more examples in our project website: https://lineargen.github.io/.
comment: 20 pages, 20 figures
☆ Deep Learning for Spectrum Prediction in Cognitive Radio Networks: State-of-the-Art, New Opportunities, and Challenges
Spectrum prediction is considered to be a promising technology that enhances spectrum efficiency by assisting dynamic spectrum access (DSA) in cognitive radio networks (CRN). Nonetheless, the highly nonlinear nature of spectrum data across time, frequency, and space domains, coupled with the intricate spectrum usage patterns, poses challenges for accurate spectrum prediction. Deep learning (DL), recognized for its capacity to extract nonlinear features, has been applied to solve these challenges. This paper first shows the advantages of applying DL by comparing with traditional prediction methods. Then, the current state-of-the-art DL-based spectrum prediction techniques are reviewed and summarized in terms of intra-band and crossband prediction. Notably, this paper uses a real-world spectrum dataset to prove the advancements of DL-based methods. Then, this paper proposes a novel intra-band spatiotemporal spectrum prediction framework named ViTransLSTM. This framework integrates visual self-attention and long short-term memory to capture both local and global long-term spatiotemporal dependencies of spectrum usage patterns. Similarly, the effectiveness of the proposed framework is validated on the aforementioned real-world dataset. Finally, the paper presents new related challenges and potential opportunities for future research.
☆ Learning Structural Causal Models from Ordering: Identifiable Flow Models AAAI 2025
In this study, we address causal inference when only observational data and a valid causal ordering from the causal graph are available. We introduce a set of flow models that can recover component-wise, invertible transformation of exogenous variables. Our flow-based methods offer flexible model design while maintaining causal consistency regardless of the number of discretization steps. We propose design improvements that enable simultaneous learning of all causal mechanisms and reduce abduction and prediction complexity to linear O(n) relative to the number of layers, independent of the number of causal variables. Empirically, we demonstrate that our method outperforms previous state-of-the-art approaches and delivers consistent performance across a wide range of structural causal models in answering observational, interventional, and counterfactual questions. Additionally, our method achieves a significant reduction in computational time compared to existing diffusion-based techniques, making it practical for large structural causal models.
comment: Accepted at AAAI 2025
☆ Precise Antigen-Antibody Structure Predictions Enhance Antibody Development with HelixFold-Multimer
The accurate prediction of antigen-antibody structures is essential for advancing immunology and therapeutic development, as it helps elucidate molecular interactions that underlie immune responses. Despite recent progress with deep learning models like AlphaFold and RoseTTAFold, accurately modeling antigen-antibody complexes remains a challenge due to their unique evolutionary characteristics. HelixFold-Multimer, a specialized model developed for this purpose, builds on the framework of AlphaFold-Multimer and demonstrates improved precision for antigen-antibody structures. HelixFold-Multimer not only surpasses other models in accuracy but also provides essential insights into antibody development, enabling more precise identification of binding sites, improved interaction prediction, and enhanced design of therapeutic antibodies. These advances underscore HelixFold-Multimer's potential in supporting antibody research and therapeutic innovation.
☆ MERaLiON-AudioLLM: Technical Report
We introduce MERaLiON-AudioLLM (Multimodal Empathetic Reasoning and Learning in One Network), the first speech-text model tailored for Singapore's multilingual and multicultural landscape. Developed under the National Large Language Models Funding Initiative, Singapore, MERaLiON-AudioLLM integrates advanced speech and text processing to address the diverse linguistic nuances of local accents and dialects, enhancing accessibility and usability in complex, multilingual environments. Our results demonstrate improvements in both speech recognition and task-specific understanding, positioning MERaLiON-AudioLLM as a pioneering solution for region specific AI applications. We envision this release to set a precedent for future models designed to address localised linguistic and cultural contexts in a global framework.
☆ Temporal Causal Discovery in Dynamic Bayesian Networks Using Federated Learning
Traditionally, learning the structure of a Dynamic Bayesian Network has been centralized, with all data pooled in one location. However, in real-world scenarios, data are often dispersed among multiple parties (e.g., companies, devices) that aim to collaboratively learn a Dynamic Bayesian Network while preserving their data privacy and security. In this study, we introduce a federated learning approach for estimating the structure of a Dynamic Bayesian Network from data distributed horizontally across different parties. We propose a distributed structure learning method that leverages continuous optimization so that only model parameters are exchanged during optimization. Experimental results on synthetic and real datasets reveal that our method outperforms other state-of-the-art techniques, particularly when there are many clients with limited individual sample sizes.
comment: 23 pages
☆ Universal Inceptive GNNs by Eliminating the Smoothness-generalization Dilemma
Graph Neural Networks (GNNs) have demonstrated remarkable success in various domains, such as transaction and social net-works. However, their application is often hindered by the varyinghomophily levels across different orders of neighboring nodes, ne-cessitating separate model designs for homophilic and heterophilicgraphs. In this paper, we aim to develop a unified framework ca-pable of handling neighborhoods of various orders and homophilylevels. Through theoretical exploration, we identify a previouslyoverlooked architectural aspect in multi-hop learning: the cascadedependency, which leads to asmoothness-generalization dilemma.This dilemma significantly affects the learning process, especiallyin the context of high-order neighborhoods and heterophilic graphs.To resolve this issue, we propose an Inceptive Graph Neural Net-work (IGNN), a universal message-passing framework that replacesthe cascade dependency with an inceptive architecture. IGNN pro-vides independent representations for each hop, allowing personal-ized generalization capabilities, and captures neighborhood-wiserelationships to select appropriate receptive fields. Extensive ex-periments show that our IGNN outperforms 23 baseline methods,demonstrating superior performance on both homophilic and het-erophilic graphs, while also scaling efficiently to large graphs.
comment: 12 pages
☆ CP-DETR: Concept Prompt Guide DETR Toward Stronger Universal Object Detection AAAI2025
Recent research on universal object detection aims to introduce language in a SoTA closed-set detector and then generalize the open-set concepts by constructing large-scale (text-region) datasets for training. However, these methods face two main challenges: (i) how to efficiently use the prior information in the prompts to genericise objects and (ii) how to reduce alignment bias in the downstream tasks, both leading to sub-optimal performance in some scenarios beyond pre-training. To address these challenges, we propose a strong universal detection foundation model called CP-DETR, which is competitive in almost all scenarios, with only one pre-training weight. Specifically, we design an efficient prompt visual hybrid encoder that enhances the information interaction between prompt and visual through scale-by-scale and multi-scale fusion modules. Then, the hybrid encoder is facilitated to fully utilize the prompted information by prompt multi-label loss and auxiliary detection head. In addition to text prompts, we have designed two practical concept prompt generation methods, visual prompt and optimized prompt, to extract abstract concepts through concrete visual examples and stably reduce alignment bias in downstream tasks. With these effective designs, CP-DETR demonstrates superior universal detection performance in a broad spectrum of scenarios. For example, our Swin-T backbone model achieves 47.6 zero-shot AP on LVIS, and the Swin-L backbone model achieves 32.2 zero-shot AP on ODinW35. Furthermore, our visual prompt generation method achieves 68.4 AP on COCO val by interactive detection, and the optimized prompt achieves 73.1 fully-shot AP on ODinW13.
comment: Accepted by AAAI2025
☆ AutoPatent: A Multi-Agent Framework for Automatic Patent Generation
As the capabilities of Large Language Models (LLMs) continue to advance, the field of patent processing has garnered increased attention within the natural language processing community. However, the majority of research has been concentrated on classification tasks, such as patent categorization and examination, or on short text generation tasks like patent summarization and patent quizzes. In this paper, we introduce a novel and practical task known as Draft2Patent, along with its corresponding D2P benchmark, which challenges LLMs to generate full-length patents averaging 17K tokens based on initial drafts. Patents present a significant challenge to LLMs due to their specialized nature, standardized terminology, and extensive length. We propose a multi-agent framework called AutoPatent which leverages the LLM-based planner agent, writer agents, and examiner agent with PGTree and RRAG to generate lengthy, intricate, and high-quality complete patent documents. The experimental results demonstrate that our AutoPatent framework significantly enhances the ability to generate comprehensive patents across various LLMs. Furthermore, we have discovered that patents generated solely with the AutoPatent framework based on the Qwen2.5-7B model outperform those produced by larger and more powerful LLMs, such as GPT-4o, Qwen2.5-72B, and LLAMA3.1-70B, in both objective metrics and human evaluations. We will make the data and code available upon acceptance at \url{https://github.com/QiYao-Wang/AutoPatent}.
comment: 19 pages, 7 figures
☆ Semi-IIN: Semi-supervised Intra-inter modal Interaction Learning Network for Multimodal Sentiment Analysis
Despite multimodal sentiment analysis being a fertile research ground that merits further investigation, current approaches take up high annotation cost and suffer from label ambiguity, non-amicable to high-quality labeled data acquisition. Furthermore, choosing the right interactions is essential because the significance of intra- or inter-modal interactions can differ among various samples. To this end, we propose Semi-IIN, a Semi-supervised Intra-inter modal Interaction learning Network for multimodal sentiment analysis. Semi-IIN integrates masked attention and gating mechanisms, enabling effective dynamic selection after independently capturing intra- and inter-modal interactive information. Combined with the self-training approach, Semi-IIN fully utilizes the knowledge learned from unlabeled data. Experimental results on two public datasets, MOSI and MOSEI, demonstrate the effectiveness of Semi-IIN, establishing a new state-of-the-art on several metrics. Code is available at https://github.com/flow-ljh/Semi-IIN.
☆ Learning Visually Grounded Domain Ontologies via Embodied Conversation and Explanation AAAI
In this paper, we offer a learning framework in which the agent's knowledge gaps are overcome through corrective feedback from a teacher whenever the agent explains its (incorrect) predictions. We test it in a low-resource visual processing scenario, in which the agent must learn to recognize distinct types of toy truck. The agent starts the learning process with no ontology about what types of trucks exist nor which parts they have, and a deficient model for recognizing those parts from visual input. The teacher's feedback to the agent's explanations addresses its lack of relevant knowledge in the ontology via a generic rule (e.g., "dump trucks have dumpers"), whereas an inaccurate part recognition is corrected by a deictic statement (e.g., "this is not a dumper"). The learner utilizes this feedback not only to improve its estimate of the hypothesis space of possible domain ontologies and probability distributions over them, but also to use those estimates to update its visual interpretation of the scene. Our experiments demonstrate that teacher-learner pairs utilizing explanations and corrections are more data-efficient than those without such a faculty.
comment: Accepted to, and to appear in the Thirty-Ninth AAAI Conference on Artificial Intelligence (AAAI-25)
♻ ☆ Inverse Reinforcement Learning by Estimating Expertise of Demonstrators AAAI
In Imitation Learning (IL), utilizing suboptimal and heterogeneous demonstrations presents a substantial challenge due to the varied nature of real-world data. However, standard IL algorithms consider these datasets as homogeneous, thereby inheriting the deficiencies of suboptimal demonstrators. Previous approaches to this issue rely on impractical assumptions like high-quality data subsets, confidence rankings, or explicit environmental knowledge. This paper introduces IRLEED, Inverse Reinforcement Learning by Estimating Expertise of Demonstrators, a novel framework that overcomes these hurdles without prior knowledge of demonstrator expertise. IRLEED enhances existing Inverse Reinforcement Learning (IRL) algorithms by combining a general model for demonstrator suboptimality to address reward bias and action variance, with a Maximum Entropy IRL framework to efficiently derive the optimal policy from diverse, suboptimal demonstrations. Experiments in both online and offline IL settings, with simulated and human-generated data, demonstrate IRLEED's adaptability and effectiveness, making it a versatile solution for learning from suboptimal demonstrations.
comment: 11 pages, 4 figures, extended version of AAAI publication
♻ ☆ DroidSpeak: KV Cache Sharing for Efficient Multi-LLM Serving
Large Language Models (LLMs) are increasingly employed in complex workflows, where different LLMs and fine-tuned variants collaboratively address complex tasks. However, these systems face significant inefficiencies due to redundant context processing of the shared context. We propose DroidSpeak, a framework that optimizes context sharing between fine-tuned LLMs derived from the same foundational model. DroidSpeak identifies critical layers in the KV cache and selectively recomputes them, enabling effective reuse of intermediate data while maintaining high accuracy. Our approach balances computational efficiency and task fidelity, significantly reducing inference latency and throughput bottlenecks. Experiments on diverse datasets and model pairs demonstrate that DroidSpeak achieves up to 3x higher throughputs and 2.6x faster prefill times with negligible accuracy loss compared to full recomputation.
♻ ☆ NLP Cluster Analysis of Common Core State Standards and NAEP Item Specifications
Camilli (2024) proposed a methodology using natural language processing (NLP) to map the relationship of a set of content standards to item specifications. This study provided evidence that NLP can be used to improve the mapping process. As part of this investigation, the nominal classifications of standards and items specifications were used to examine construct equivalence. In the current paper, we determine the strength of empirical support for the semantic distinctiveness of these classifications, which are known as "domains" for Common Core standards, and "strands" for National Assessment of Educational Progress (NAEP) item specifications. This is accomplished by separate k-means clustering for standards and specifications of their corresponding embedding vectors. We then briefly illustrate an application of these findings.
comment: 10 pages, 5 tables
♻ ☆ Solving Epistemic Logic Programs using Generate-and-Test with Propagation AAAI
This paper introduces a general framework for generate-and-test-based solvers for epistemic logic programs that can be instantiated with different generator and tester programs, and we prove sufficient conditions on those programs for the correctness of the solvers built using this framework. It also introduces a new generator program that incorporates the propagation of epistemic consequences and shows that this can exponentially reduce the number of candidates that need to be tested while only incurring a linear overhead. We implement a new solver based on these theoretical findings and experimentally show that it outperforms existing solvers by achieving a ~3.3x speed-up and solving 91% more instances on well-known benchmarks.
comment: Accepted for publication in the Proceedings of the 39th Annual AAAI Conference on Artificial Intelligence
♻ ☆ Fine Tuning Large Language Models for Medicine: The Role and Importance of Direct Preference Optimization
Large Language Model (LLM) fine tuning is underutilized in the field of medicine. Two of the most common methods of fine tuning are Supervised Fine Tuning (SFT) and Direct Preference Optimization (DPO), but there is little guidance informing users when to use either technique. In this investigation, we compare the performance of SFT and DPO for five common natural language tasks in medicine: Classification with text data, Classification with numeric data, Clinical Reasoning, Summarization, and Clinical Triage. We find that SFT alone is sufficient for Classification with text data, whereas DPO improves performance for the more complex tasks of Clinical Reasoning, Summarization and Clinical Triage. Our results establish the role and importance of DPO fine tuning within medicine, and consequently call attention to current software gaps that prevent widespread deployment of this technique.
♻ ☆ A logic for reasoning with inconsistent knowledge -- A reformulation using nowadays terminology (2024)
In many situations humans have to reason with inconsistent knowledge. These inconsistencies may occur due to not fully reliable sources of information. In order to reason with inconsistent knowledge, it is not possible to view a set of premisses as absolute truths as is done in predicate logic. Viewing the set of premisses as a set of assumptions, however, it is possible to deduce useful conclusions from an inconsistent set of premisses. In this paper a logic for reasoning with inconsistent knowledge is described. This logic is a generalization of the work of N. Rescher [15]. In the logic a reliability relation is used to choose between incompatible assumptions. These choices are only made when a contradiction is derived. As long as no contradiction is derived, the knowledge is assumed to be consistent. This makes it possible to define an argumentation-based deduction process for the logic. For the logic a semantics based on the ideas of Y. Shoham [22, 23], is defined. It turns out that the semantics for the logic is a preferential semantics according to the definition S. Kraus, D. Lehmann and M. Magidor [12]. Therefore the logic is a logic of system P and possesses all the properties of an ideal non-monotonic logic.
comment: The original version was published in the Artificial Intelligence journal. This original version uses 'justifications' in the proof system, which we would call nowadays 'arguments'. The current version presents the same results but now using the terminology of an assumption-based argumentation system
♻ ☆ TrustUQA: A Trustful Framework for Unified Structured Data Question Answering AAAI 2025
Natural language question answering (QA) over structured data sources such as tables and knowledge graphs have been widely investigated, especially with Large Language Models (LLMs) in recent years. The main solutions include question to formal query parsing and retrieval-based answer generation. However, current methods of the former often suffer from weak generalization, failing to dealing with multi-types of sources, while the later is limited in trustfulness. In this paper, we propose TrustUQA, a trustful QA framework that can simultaneously support multiple types of structured data in a unified way. To this end, it adopts an LLM-friendly and unified knowledge representation method called Condition Graph(CG), and uses an LLM and demonstration-based two-level method for CG querying. For enhancement, it is also equipped with dynamic demonstration retrieval. We have evaluated TrustUQA with 5 benchmarks covering 3 types of structured data. It outperforms 2 existing unified structured data QA methods. In comparison with the baselines that are specific to one data type, it achieves state-of-the-art on 2 of the datasets. Further more, we have demonstrated the potential of our method for more general QA tasks, QA over mixed structured data and QA across structured data. The code is available at https://github.com/zjukg/TrustUQA.
comment: Accepted by AAAI 2025
♻ ☆ SoK: Decentralized AI (DeAI)
The centralization of Artificial Intelligence (AI) poses significant challenges, including single points of failure, inherent biases, data privacy concerns, and scalability issues. These problems are especially prevalent in closed-source large language models (LLMs), where user data is collected and used without transparency. To mitigate these issues, blockchain-based decentralized AI (DeAI) has emerged as a promising solution. DeAI combines the strengths of both blockchain and AI technologies to enhance the transparency, security, decentralization, and trustworthiness of AI systems. However, a comprehensive understanding of state-of-the-art DeAI development, particularly for active industry solutions, is still lacking. In this work, we present a Systematization of Knowledge (SoK) for blockchain-based DeAI solutions. We propose a taxonomy to classify existing DeAI protocols based on the model lifecycle. Based on this taxonomy, we provide a structured way to clarify the landscape of DeAI protocols and identify their similarities and differences. We analyze the functionalities of blockchain in DeAI, investigating how blockchain features contribute to enhancing the security, transparency, and trustworthiness of AI processes, while also ensuring fair incentives for AI data and model contributors. In addition, we identify key insights and research gaps in developing DeAI protocols, highlighting several critical avenues for future research.
comment: This is a Systematization of Knowledge (SoK) for the rapidly evolving field of Decentralized AI (DeAI). We welcome valuable comments, suggestions, and collaboration to further refine and enhance this work. We hope our contribution will help accelerate the advancement of DeAI
♻ ☆ Robust Monocular Visual Odometry using Curriculum Learning
Curriculum Learning (CL), drawing inspiration from natural learning patterns observed in humans and animals, employs a systematic approach of gradually introducing increasingly complex training data during model development. Our work applies innovative CL methodologies to address the challenging geometric problem of monocular Visual Odometry (VO) estimation, which is essential for robot navigation in constrained environments. The primary objective of our research is to push the boundaries of current state-of-the-art (SOTA) benchmarks in monocular VO by investigating various curriculum learning strategies. We enhance the end-to-end Deep-Patch-Visual Odometry (DPVO) framework through the integration of novel CL approaches, with the goal of developing more resilient models capable of maintaining high performance across challenging environments and complex motion scenarios. Our research encompasses several distinctive CL strategies. We develop methods to evaluate sample difficulty based on trajectory motion characteristics, implement sophisticated adaptive scheduling through self-paced weighted loss mechanisms, and utilize reinforcement learning agents for dynamic adjustment of training emphasis. Through comprehensive evaluation on the diverse synthetic TartanAir dataset and complex real-world benchmarks such as EuRoC and TUM-RGBD, our Curriculum Learning-based Deep-Patch-Visual Odometry (CL-DPVO) demonstrates superior performance compared to existing SOTA methods, including both feature-based and learning-based VO approaches. The results validate the effectiveness of integrating curriculum learning principles into visual odometry systems.
comment: 11 pages
♻ ☆ IGNITE: Individualized GeNeration of Imputations in Time-series Electronic health records
Electronic Health Records present a valuable modality for driving personalized medicine, where treatment is tailored to fit individual-level differences. For this purpose, many data-driven machine learning and statistical models rely on the wealth of longitudinal EHRs to study patients' physiological and treatment effects. However, longitudinal EHRs tend to be sparse and highly missing, where missingness could also be informative and reflect the underlying patient's health status. Therefore, the success of data-driven models for personalized medicine highly depends on how the EHR data is represented from physiological data, treatments, and the missing values in the data. To this end, we propose a novel deep-learning model that learns the underlying patient dynamics over time across multivariate data to generate personalized realistic values conditioning on an individual's demographic characteristics and treatments. Our proposed model, IGNITE (Individualized GeNeration of Imputations in Time-series Electronic health records), utilises a conditional dual-variational autoencoder augmented with dual-stage attention to generate missing values for an individual. In IGNITE, we further propose a novel individualized missingness mask (IMM), which helps our model generate values based on the individual's observed data and missingness patterns. We further extend the use of IGNITE from imputing missingness to a personalized data synthesizer, where it generates missing EHRs that were never observed prior or even generates new patients for various applications. We validate our model on three large publicly available datasets and show that IGNITE outperforms state-of-the-art approaches in missing data reconstruction and task prediction.
♻ ☆ Pre-Deployment Information Sharing: A Zoning Taxonomy for Precursory Capabilities
High-impact and potentially dangerous capabilities can and should be broken down into early warning shots long before reaching red lines. Each of these early warning shots should correspond to a precursory capability. Each precursory capability sits on a spectrum indicating its proximity to a final high-impact capability, corresponding to a red line. To meaningfully detect and track capability progress, we propose a taxonomy of dangerous capability zones (a zoning taxonomy) tied to a staggered information exchange framework that enables relevant bodies to take action accordingly. In the Frontier AI Safety Commitments, signatories commit to sharing more detailed information with trusted actors, including an appointed body, as appropriate (Commitment VII). Building on our zoning taxonomy, this paper makes four recommendations for specifying information sharing as detailed in Commitment VII. (1) Precursory capabilities should be shared as soon as they become known through internal evaluations before deployment. (2) AI Safety Institutes (AISIs) should be the trusted actors appointed to receive and coordinate information on precursory components. (3) AISIs should establish adequate information protection infrastructure and guarantee increased information security as precursory capabilities move through the zones and towards red lines, including, if necessary, by classifying the information on precursory capabilities or marking it as controlled. (4) High-impact capability progress in one geographical region may translate to risk in other regions and necessitates more comprehensive risk assessment internationally. As such, AISIs should exchange information on precursory capabilities with other AISIs, relying on the existing frameworks on international classified exchanges and applying lessons learned from other regulated high-risk sectors.
♻ ☆ Olympus: A Universal Task Router for Computer Vision Tasks
We introduce Olympus, a new approach that transforms Multimodal Large Language Models (MLLMs) into a unified framework capable of handling a wide array of computer vision tasks. Utilizing a controller MLLM, Olympus delegates over 20 specialized tasks across images, videos, and 3D objects to dedicated modules. This instruction-based routing enables complex workflows through chained actions without the need for training heavy generative models. Olympus easily integrates with existing MLLMs, expanding their capabilities with comparable performance. Experimental results demonstrate that Olympus achieves an average routing accuracy of 94.75% across 20 tasks and precision of 91.82% in chained action scenarios, showcasing its effectiveness as a universal task router that can solve a diverse range of computer vision tasks. Project page: http://yuanze-lin.me/Olympus_page/
comment: Technical Report
♻ ☆ Structural Entropy Guided Probabilistic Coding AAAI 2025
Probabilistic embeddings have several advantages over deterministic embeddings as they map each data point to a distribution, which better describes the uncertainty and complexity of data. Many works focus on adjusting the distribution constraint under the Information Bottleneck (IB) principle to enhance representation learning. However, these proposed regularization terms only consider the constraint of each latent variable, omitting the structural information between latent variables. In this paper, we propose a novel structural entropy-guided probabilistic coding model, named SEPC. Specifically, we incorporate the relationship between latent variables into the optimization by proposing a structural entropy regularization loss. Besides, as traditional structural information theory is not well-suited for regression tasks, we propose a probabilistic encoding tree, transferring regression tasks to classification tasks while diminishing the influence of the transformation. Experimental results across 12 natural language understanding tasks, including both classification and regression tasks, demonstrate the superior performance of SEPC compared to other state-of-the-art models in terms of effectiveness, generalization capability, and robustness to label noise. The codes and datasets are available at https://github.com/SELGroup/SEPC.
comment: This paper is accepted by AAAI 2025
♻ ☆ Symmetric masking strategy enhances the performance of Masked Image Modeling
Masked Image Modeling (MIM) is a technique in self-supervised learning that focuses on acquiring detailed visual representations from unlabeled images by estimating the missing pixels in randomly masked sections. It has proven to be a powerful tool for the preliminary training of Vision Transformers (ViTs), yielding impressive results across various tasks. Nevertheless, most MIM methods heavily depend on the random masking strategy to formulate the pretext task. This strategy necessitates numerous trials to ascertain the optimal dropping ratio, which can be resource-intensive, requiring the model to be pre-trained for anywhere between 800 to 1600 epochs. Furthermore, this approach may not be suitable for all datasets. In this work, we propose a new masking strategy that effectively helps the model capture global and local features. Based on this masking strategy, SymMIM, our proposed training pipeline for MIM is introduced. SymMIM achieves a new SOTA accuracy of 85.9\% on ImageNet using ViT-Large and surpasses previous SOTA across downstream tasks such as image classification, semantic segmentation, object detection, instance segmentation tasks, and so on.
♻ ☆ Towards Cross-Lingual Audio Abuse Detection in Low-Resource Settings with Few-Shot Learning COLING 2025
Online abusive content detection, particularly in low-resource settings and within the audio modality, remains underexplored. We investigate the potential of pre-trained audio representations for detecting abusive language in low-resource languages, in this case, in Indian languages using Few Shot Learning (FSL). Leveraging powerful representations from models such as Wav2Vec and Whisper, we explore cross-lingual abuse detection using the ADIMA dataset with FSL. Our approach integrates these representations within the Model-Agnostic Meta-Learning (MAML) framework to classify abusive language in 10 languages. We experiment with various shot sizes (50-200) evaluating the impact of limited data on performance. Additionally, a feature visualization study was conducted to better understand model behaviour. This study highlights the generalization ability of pre-trained models in low-resource scenarios and offers valuable insights into detecting abusive language in multilingual contexts.
comment: Accepted as part of the proceedings of COLING 2025
♻ ☆ SVGDreamer++: Advancing Editability and Diversity in Text-Guided SVG Generation
Recently, text-guided scalable vector graphics (SVG) synthesis has demonstrated significant potential in domains such as iconography and sketching. However, SVGs generated from existing Text-to-SVG methods often lack editability and exhibit deficiencies in visual quality and diversity. In this paper, we propose a novel text-guided vector graphics synthesis method to address these limitations. To enhance the editability of output SVGs, we introduce a Hierarchical Image VEctorization (HIVE) framework that operates at the semantic object level and supervises the optimization of components within the vector object. This approach facilitates the decoupling of vector graphics into distinct objects and component levels. Our proposed HIVE algorithm, informed by image segmentation priors, not only ensures a more precise representation of vector graphics but also enables fine-grained editing capabilities within vector objects. To improve the diversity of output SVGs, we present a Vectorized Particle-based Score Distillation (VPSD) approach. VPSD addresses over-saturation issues in existing methods and enhances sample diversity. A pre-trained reward model is incorporated to re-weight vector particles, improving aesthetic appeal and enabling faster convergence. Additionally, we design a novel adaptive vector primitives control strategy, which allows for the dynamic adjustment of the number of primitives, thereby enhancing the presentation of graphic details. Extensive experiments validate the effectiveness of the proposed method, demonstrating its superiority over baseline methods in terms of editability, visual quality, and diversity. We also show that our new method supports up to six distinct vector styles, capable of generating high-quality vector assets suitable for stylized vector design and poster design. Code and demo will be released at: http://ximinng.github.io/SVGDreamerV2Project/
comment: 17 pages, 17 figures. Project Page: http://ximinng.github.io/SVGDreamerV2Project/. arXiv admin note: text overlap with arXiv:2312.16476
♻ ☆ GATEAU: Selecting Influential Sample for Long Context Alignment
Aligning large language models to handle instructions with extremely long contexts has yet to be fully investigated. Previous studies attempt to scale up the available data volume by synthesizing long instruction-following samples, as constructing such a dataset tends to be challenging for annotators. However, a lack of a well-defined strategy for ensuring data quality may introduce low-quality samples and restrict the model performance. Thus, we propose GATEAU, a novel framework to address the unique challenge of long context alignment by identifying the influential samples enriched with long-range dependency relations. Specifically, GATEAU measures the long-range dependencies from two essential aspects: the difficulty of generating target responses due to the long-range dependencies, and the difficulty of understanding long inputs due to such dependencies. Comprehensive experiments indicate that GATEAU effectively identifies influential samples and the model trained on these selected samples exhibits better instruction-following and long-context understanding capabilities.
♻ ☆ AgentMixer: Multi-Agent Correlated Policy Factorization
In multi-agent reinforcement learning, centralized training with decentralized execution (CTDE) methods typically assume that agents make decisions based on their local observations independently, which may not lead to a correlated joint policy with coordination. Coordination can be explicitly encouraged during training and individual policies can be trained to imitate the correlated joint policy. However, this may lead to an \textit{asymmetric learning failure} due to the observation mismatch between the joint and individual policies. Inspired by the concept of correlated equilibrium, we introduce a \textit{strategy modification} called AgentMixer that allows agents to correlate their policies. AgentMixer combines individual partially observable policies into a joint fully observable policy non-linearly. To enable decentralized execution, we introduce \textit{Individual-Global-Consistency} to guarantee mode consistency during joint training of the centralized and decentralized policies and prove that AgentMixer converges to an $\epsilon$-approximate Correlated Equilibrium. In the Multi-Agent MuJoCo, SMAC-v2, Matrix Game, and Predator-Prey benchmarks, AgentMixer outperforms or matches state-of-the-art methods.
♻ ☆ TreeEval: Benchmark-Free Evaluation of Large Language Models through Tree Planning
Recently, numerous new benchmarks have been established to evaluate the performance of large language models (LLMs) via either computing a holistic score or employing another LLM as a judge. However, these approaches suffer from data leakage due to the open access of the benchmark and inflexible evaluation process. To address this issue, we introduce $\textbf{TreeEval}$, a benchmark-free evaluation method for LLMs that let a high-performance LLM host an irreproducible evaluation session and essentially avoids the data leakage. Moreover, this LLM performs as an examiner to raise up a series of questions under a topic with a tree planing strategy, which considers the current evaluation status to decide the next question generation and ensures the completeness and efficiency of the evaluation process. We evaluate $6$ models of different parameter sizes, including $7$B, $13$B, and $33$B, and ultimately achieved the highest correlation coefficient with AlpacaEval2.0 using only around $45$ questions. We also conduct more analysis to show the robustness and reliability of TreeEval. Our code can be accessed via the provided https://github.com/Ashura5/TreeEval.
♻ ☆ Unveiling the optimization process of Physics Informed Neural Networks: How accurate and competitive can PINNs be?
This study investigates the potential accuracy boundaries of physics-informed neural networks, contrasting their approach with previous similar works and traditional numerical methods. We find that selecting improved optimization algorithms significantly enhances the accuracy of the results. Simple modifications to the loss function may also improve precision, offering an additional avenue for enhancement. Despite optimization algorithms having a greater impact on convergence than adjustments to the loss function, practical considerations often favor tweaking the latter due to ease of implementation. On a global scale, the integration of an enhanced optimizer and a marginally adjusted loss function enables a reduction in the loss function by several orders of magnitude across diverse physical problems. Consequently, our results obtained using compact networks (typically comprising 2 or 3 layers of 20-30 neurons) achieve accuracies comparable to finite difference schemes employing thousands of grid points. This study encourages the continued advancement of PINNs and associated optimization techniques for broader applications across various fields.
comment: 63 pages, 25 figures. This is the author-accepted manuscript of the paper published in Journal of Computational Physics
♻ ☆ Hidden Biases of End-to-End Driving Datasets CVPR 2024
End-to-end driving systems have made rapid progress, but have so far not been applied to the challenging new CARLA Leaderboard 2.0. Further, while there is a large body of literature on end-to-end architectures and training strategies, the impact of the training dataset is often overlooked. In this work, we make a first attempt at end-to-end driving for Leaderboard 2.0. Instead of investigating architectures, we systematically analyze the training dataset, leading to new insights: (1) Expert style significantly affects downstream policy performance. (2) In complex data sets, the frames should not be weighted on the basis of simplistic criteria such as class frequencies. (3) Instead, estimating whether a frame changes the target labels compared to previous frames can reduce the size of the dataset without removing important information. By incorporating these findings, our model ranks first and second respectively on the map and sensors tracks of the 2024 CARLA Challenge, and sets a new state-of-the-art on the Bench2Drive test routes. Finally, we uncover a design flaw in the current evaluation metrics and propose a modification for future challenges. Our dataset, code, and pre-trained models are publicly available at https://github.com/autonomousvision/carla_garage.
comment: Technical report for the CVPR 2024 Workshop on Foundation Models for Autonomous Systems. Runner-up of the track 'CARLA Autonomous Driving Challenge' in the 2024 Autonomous Grand Challenge (https://opendrivelab.com/challenge2024/)
♻ ☆ Uncertainty-Aware Artificial Intelligence for Gear Fault Diagnosis in Motor Drives
This paper introduces a novel approach to quantify the uncertainties in fault diagnosis of motor drives using Bayesian neural networks (BNN). Conventional data-driven approaches used for fault diagnosis often rely on point-estimate neural networks, which merely provide deterministic outputs and fail to capture the uncertainty associated with the inference process. In contrast, BNNs offer a principled framework to model uncertainty by treating network weights as probability distributions rather than fixed values. It offers several advantages: (a) improved robustness to noisy data, (b) enhanced interpretability of model predictions, and (c) the ability to quantify uncertainty in the decision-making processes. To test the robustness of the proposed BNN, it has been tested under a conservative dataset of gear fault data from an experimental prototype of three fault types at first, and is then incrementally trained on new fault classes and datasets to explore its uncertainty quantification features and model interpretability under noisy data and unseen fault scenarios.
comment: The manuscript has been accepted for publication in 2025 IEEE Applied Power Electronics Conference and Exposition (APEC)
♻ ☆ HyViLM: Enhancing Fine-Grained Recognition with a Hybrid Encoder for Vision-Language Models
Recently, there has been growing interest in the capability of multimodal large language models (MLLMs) to process high-resolution images. A common approach currently involves dynamically cropping the original high-resolution image into smaller sub-images, which are then fed into a vision encoder that was pre-trained on lower-resolution images. However, this cropping approach often truncates objects and connected areas in the original image, causing semantic breaks. To address this limitation, we introduce HyViLM, designed to process images of any resolution while retaining the overall context during encoding. Specifically, we: (i) Design a new visual encoder called Hybrid Encoder that not only encodes individual sub-images but also interacts with detailed global visual features, significantly improving the model's ability to encode high-resolution images. (ii) Propose an optimal feature fusion strategy for the dynamic cropping approach, effectively leveraging information from different layers of the vision encoder. Compared with the state-of-the-art MLLMs under the same setting, our HyViLM outperforms existing MLLMs in nine out of ten tasks. Specifically, HyViLM achieves a 9.6% improvement in performance on the TextVQA task and a 6.9% enhancement on the DocVQA task.
comment: 11 pages, 4 figures
♻ ☆ Benchmarking LLMs for Mimicking Child-Caregiver Language in Interaction
LLMs can generate human-like dialogues, yet their ability to simulate early child-adult interactions remains largely unexplored. In this paper, we examined how effectively LLMs can capture the distinctive features of child-caregiver language in interaction, using both static and interactive benchmarking methods. We found that state-of-the-art LLMs like Llama 3 and GPT-4o can approximate child-caregiver dialogues at the word and utterance level, but they struggle to reproduce the child and caregiver's discursive patterns, exaggerate alignment, and fail to reach the level of diversity shown by humans. The broader goal of this work is to initiate the development of a comprehensive benchmark for LLMs in child-oriented applications.
♻ ☆ Dynamic Fog Computing for Enhanced LLM Execution in Medical Applications
The ability of large language models (LLMs) to transform, interpret, and comprehend vast quantities of heterogeneous data presents a significant opportunity to enhance data-driven care delivery. However, the sensitive nature of protected health information (PHI) raises valid concerns about data privacy and trust in remote LLM platforms. In addition, the cost associated with cloud-based artificial intelligence (AI) services continues to impede widespread adoption. To address these challenges, we propose a shift in the LLM execution environment from opaque, centralized cloud providers to a decentralized and dynamic fog computing architecture. By executing open-weight LLMs in more trusted environments, such as the user's edge device or a fog layer within a local network, we aim to mitigate the privacy, trust, and financial challenges associated with cloud-based LLMs. We further present SpeziLLM, an open-source framework designed to facilitate rapid and seamless leveraging of different LLM execution layers and lowering barriers to LLM integration in digital health applications. We demonstrate SpeziLLM's broad applicability across six digital health applications, showcasing its versatility in various healthcare settings.
♻ ☆ Solid Waste Detection, Monitoring and Mapping in Remote Sensing Images: A Survey
The detection and characterization of illegal solid waste disposal sites are essential for environmental protection, particularly for mitigating pollution and health hazards. Improperly managed landfills contaminate soil and groundwater via rainwater infiltration, posing threats to both animals and humans. Traditional landfill identification approaches, such as on-site inspections, are time-consuming and expensive. Remote sensing is a cost-effective solution for the identification and monitoring of solid waste disposal sites that enables broad coverage and repeated acquisitions over time. Earth Observation (EO) satellites, equipped with an array of sensors and imaging capabilities, have been providing high-resolution data for several decades. Researchers proposed specialized techniques that leverage remote sensing imagery to perform a range of tasks such as waste site detection, dumping site monitoring, and assessment of suitable locations for new landfills. This review aims to provide a detailed illustration of the most relevant proposals for the detection and monitoring of solid waste sites by describing and comparing the approaches, the implemented techniques, and the employed data. Furthermore, since the data sources are of the utmost importance for developing an effective solid waste detection model, a comprehensive overview of the satellites and publicly available data sets is presented. Finally, this paper identifies the open issues in the state-of-the-art and discusses the relevant research directions for reducing the costs and improving the effectiveness of novel solid waste detection methods.
♻ ☆ Towards Automated Cross-domain Exploratory Data Analysis through Large Language Models SIGMOD 2025
Exploratory data analysis (EDA), coupled with SQL, is essential for data analysts involved in data exploration and analysis. However, data analysts often encounter two primary challenges: (1) the need to craft SQL queries skillfully, and (2) the requirement to generate suitable visualization types that enhance the interpretation of query results. Due to its significance, substantial research efforts have been made to explore different approaches to address these challenges, including leveraging large language models (LLMs). However, existing methods fail to meet real-world data exploration requirements primarily due to (1) complex database schema; (2) unclear user intent; (3) limited cross-domain generalization capability; and (4) insufficient end-to-end text-to-visualization capability. This paper presents TiInsight, an automated SQL-based cross-domain exploratory data analysis system. First, we propose hierarchical data context (i.e., HDC), which leverages LLMs to summarize the contexts related to the database schema, which is crucial for open-world EDA systems to generalize across data domains. Second, the EDA system is divided into four components (i.e., stages): HDC generation, question clarification and decomposition, text-to-SQL generation (i.e., TiSQL), and data visualization (i.e., TiChart). Finally, we implemented an end-to-end EDA system with a user-friendly GUI interface in the production environment at PingCAP. We have also open-sourced all APIs of TiInsight to facilitate research within the EDA community. Through extensive evaluations by a real-world user study, we demonstrate that TiInsight offers remarkable performance compared to human experts. Specifically, TiSQL achieves an execution accuracy of 86.3% on the Spider dataset using GPT-4. It also demonstrates state-of-the-art performance on the Bird dataset.
comment: 14 pages, 10 figures. Submitted to SIGMOD 2025
♻ ☆ AtomThink: A Slow Thinking Framework for Multimodal Mathematical Reasoning
In this paper, we address the challenging task of multimodal mathematical reasoning by incorporating the ability of ``slow thinking" into multimodal large language models (MLLMs). Contrary to existing methods that rely on direct or fast thinking, our key idea is to construct long chains of thought (CoT) consisting of atomic actions in a step-by-step manner, guiding MLLMs to perform complex reasoning. To this end, we design a novel AtomThink framework composed of three key modules: (i) a CoT annotation engine that automatically generates high-quality CoT annotations to address the lack of high-quality visual mathematical data; (ii) an atomic step fine-tuning strategy that jointly optimizes an MLLM and a policy reward model (PRM) for step-wise reasoning; and (iii) four different search strategies that can be applied with the PRM to complete reasoning. Additionally, we propose AtomMATH, a large-scale multimodal dataset of long CoTs, and an atomic capability evaluation metric for mathematical tasks. Extensive experimental results show that the proposed AtomThink significantly improves the performance of baseline MLLMs, achieving approximately 50\% relative accuracy gains on MathVista and 120\% on MathVerse. To support the advancement of multimodal slow-thinking models, we will make our code and dataset publicly available on https://github.com/Quinn777/AtomThink.
♻ ☆ Navigating the Future of Federated Recommendation Systems with Foundation Models
In recent years, the integration of federated learning (FL) and recommendation systems (RS), known as Federated Recommendation Systems (FRS), has attracted attention for preserving user privacy by keeping private data on client devices. However, FRS faces inherent limitations such as data heterogeneity and scarcity, due to the privacy requirements of FL and the typical data sparsity issues of RSs. Models like ChatGPT are empowered by the concept of transfer learning and self-supervised learning, so they can be easily applied to the downstream tasks after fine-tuning or prompting. These models, so-called Foundation Models (FM), fouce on understanding the human's intent and perform following their designed roles in the specific tasks, which are widely recognized for producing high-quality content in the image and language domains. Thus, the achievements of FMs inspire the design of FRS and suggest a promising research direction: integrating foundation models to address the above limitations. In this study, we conduct a comprehensive review of FRSs with FMs. Specifically, we: 1) summarise the common approaches of current FRSs and FMs; 2) review the challenges posed by FRSs and FMs; 3) discuss potential future research directions; and 4) introduce some common benchmarks and evaluation metrics in the FRS field. We hope that this position paper provides the necessary background and guidance to explore this interesting and emerging topic.
comment: 20 pages, position paper, survey
♻ ☆ First Train to Generate, then Generate to Train: UnitedSynT5 for Few-Shot NLI
Natural Language Inference (NLI) tasks require identifying the relationship between sentence pairs, typically classified as entailment, contradiction, or neutrality. While the current state-of-the-art (SOTA) model, Entailment Few-Shot Learning (EFL), achieves a 93.1% accuracy on the Stanford Natural Language Inference (SNLI) dataset, further advancements are constrained by the dataset's limitations. To address this, we propose a novel approach leveraging synthetic data augmentation to enhance dataset diversity and complexity. We present UnitedSynT5, an advanced extension of EFL that leverages a T5-based generator to synthesize additional premise-hypothesis pairs, which are rigorously cleaned and integrated into the training data. These augmented examples are processed within the EFL framework, embedding labels directly into hypotheses for consistency. We train a GTR-T5-XL model on this expanded dataset, achieving a new benchmark of 94.7% accuracy on the SNLI dataset, 94.0% accuracy on the E-SNLI dataset, and 92.6% accuracy on the MultiNLI dataset, surpassing the previous SOTA models. This research demonstrates the potential of synthetic data augmentation in improving NLI models, offering a path forward for further advancements in natural language understanding tasks.
comment: 14 pages
♻ ☆ A Survey of Mamba
As one of the most representative DL techniques, Transformer architecture has empowered numerous advanced models, especially the large language models (LLMs) that comprise billions of parameters, becoming a cornerstone in deep learning. Despite the impressive achievements, Transformers still face inherent limitations, particularly the time-consuming inference resulting from the quadratic computation complexity of attention calculation. Recently, a novel architecture named Mamba, drawing inspiration from classical state space models (SSMs), has emerged as a promising alternative for building foundation models, delivering comparable modeling abilities to Transformers while preserving near-linear scalability concerning sequence length. This has sparked an increasing number of studies actively exploring Mamba's potential to achieve impressive performance across diverse domains. Given such rapid evolution, there is a critical need for a systematic review that consolidates existing Mamba-empowered models, offering a comprehensive understanding of this emerging model architecture. In this survey, we therefore conduct an in-depth investigation of recent Mamba-associated studies, covering three main aspects: the advancements of Mamba-based models, the techniques of adapting Mamba to diverse data, and the applications where Mamba can excel. Specifically, we first review the foundational knowledge of various representative deep learning models and the details of Mamba-1&2 as preliminaries. Then, to showcase the significance of Mamba for AI, we comprehensively review the related studies focusing on Mamba models' architecture design, data adaptability, and applications. Finally, we present a discussion of current limitations and explore various promising research directions to provide deeper insights for future investigations.
♻ ☆ DNN Task Assignment in UAV Networks: A Generative AI Enhanced Multi-Agent Reinforcement Learning Approach
Unmanned Aerial Vehicles (UAVs) possess high mobility and flexible deployment capabilities, prompting the development of UAVs for various application scenarios within the Internet of Things (IoT). The unique capabilities of UAVs give rise to increasingly critical and complex tasks in uncertain and potentially harsh environments. The substantial amount of data generated from these applications necessitates processing and analysis through deep neural networks (DNNs). However, UAVs encounter challenges due to their limited computing resources when managing DNN models. This paper presents a joint approach that combines multiple-agent reinforcement learning (MARL) and generative diffusion models (GDM) for assigning DNN tasks to a UAV swarm, aimed at reducing latency from task capture to result output. To address these challenges, we first consider the task size of the target area to be inspected and the shortest flying path as optimization constraints, employing a greedy algorithm to resolve the subproblem with a focus on minimizing the UAV's flying path and the overall system cost. In the second stage, we introduce a novel DNN task assignment algorithm, termed GDM-MADDPG, which utilizes the reverse denoising process of GDM to replace the actor network in multi-agent deep deterministic policy gradient (MADDPG). This approach generates specific DNN task assignment actions based on agents' observations in a dynamic environment. Simulation results indicate that our algorithm performs favorably compared to benchmarks in terms of path planning, Age of Information (AoI), energy consumption, and task load balancing.
♻ ☆ See Where You Read with Eye Gaze Tracking and Large Language Model
Losing track of reading progress during line switching can be frustrating. Eye gaze tracking technology offers a potential solution by highlighting read paragraphs, aiding users in avoiding wrong line switches. However, the gap between gaze tracking accuracy (2-3 cm) and text line spacing (3-5 mm) makes direct application impractical. Existing methods leverage the linear reading pattern but fail during jump reading. This paper presents a reading tracking and highlighting system that supports both linear and jump reading. Based on experimental insights from the gaze nature study of 16 users, two gaze error models are designed to enable both jump reading detection and relocation. The system further leverages the large language model's contextual perception capability in aiding reading tracking. A reading tracking domain-specific line-gaze alignment opportunity is also exploited to enable dynamic and frequent calibration of the gaze results. Controlled experiments demonstrate reliable linear reading tracking, as well as 84% accuracy in tracking jump reading. Furthermore, real field tests with 18 volunteers demonstrated the system's effectiveness in tracking and highlighting read paragraphs, improving reading efficiency, and enhancing user experience.
comment: 9 pages
♻ ☆ Leveraging Large Language Models to Detect npm Malicious Packages
Existing malicious code detection techniques demand the integration of multiple tools to detect different malware patterns, often suffering from high misclassification rates. Therefore, malicious code detection techniques could be enhanced by adopting advanced, more automated approaches to achieve high accuracy and a low misclassification rate. The goal of this study is to aid security analysts in detecting malicious packages by empirically studying the effectiveness of Large Language Models (LLMs) in detecting malicious code. We present SocketAI, a malicious code review workflow to detect malicious code. To evaluate the effectiveness of SocketAI, we leverage a benchmark dataset of 5,115 npm packages, of which 2,180 packages have malicious code. We conducted a baseline comparison of GPT-3 and GPT-4 models with the state-of-the-art CodeQL static analysis tool, using 39 custom CodeQL rules developed in prior research to detect malicious Javascript code. We also compare the effectiveness of static analysis as a pre-screener with SocketAI workflow, measuring the number of files that need to be analyzed. and the associated costs. Additionally, we performed a qualitative study to understand the types of malicious activities detected or missed by our workflow. Our baseline comparison demonstrates a 16% and 9% improvement over static analysis in precision and F1 scores, respectively. GPT-4 achieves higher accuracy with 99% precision and 97% F1 scores, while GPT-3 offers a more cost-effective balance at 91% precision and 94% F1 scores. Pre-screening files with a static analyzer reduces the number of files requiring LLM analysis by 77.9% and decreases costs by 60.9% for GPT-3 and 76.1% for GPT-4. Our qualitative analysis identified data theft, suspicious domain connection, and arbitrary code execution as the top detected malicious activities.
comment: 13 pages, 2 Figure, 6 tables
♻ ☆ Political Actor Agent: Simulating Legislative System for Roll Call Votes Prediction with Large Language Models AAAI 2025
Predicting roll call votes through modeling political actors has emerged as a focus in quantitative political science and computer science. Widely used embedding-based methods generate vectors for legislators from diverse data sets to predict legislative behaviors. However, these methods often contend with challenges such as the need for manually predefined features, reliance on extensive training data, and a lack of interpretability. Achieving more interpretable predictions under flexible conditions remains an unresolved issue. This paper introduces the Political Actor Agent (PAA), a novel agent-based framework that utilizes Large Language Models to overcome these limitations. By employing role-playing architectures and simulating legislative system, PAA provides a scalable and interpretable paradigm for predicting roll-call votes. Our approach not only enhances the accuracy of predictions but also offers multi-view, human-understandable decision reasoning, providing new insights into political actor behaviors. We conducted comprehensive experiments using voting records from the 117-118th U.S. House of Representatives, validating the superior performance and interpretability of PAA. This study not only demonstrates PAA's effectiveness but also its potential in political science research.
comment: Accepted at AAAI 2025
♻ ☆ SMMF: Square-Matricized Momentum Factorization for Memory-Efficient Optimization
We propose SMMF (Square-Matricized Momentum Factorization), a memory-efficient optimizer that reduces the memory requirement of the widely used adaptive learning rate optimizers, such as Adam, by up to 96%. SMMF enables flexible and efficient factorization of an arbitrary rank (shape) of the first and second momentum tensors during optimization, based on the proposed square-matricization and one-time single matrix factorization. From this, it becomes effectively applicable to any rank (shape) of momentum tensors, i.e., bias, matrix, and any rank-d tensors, prevalent in various deep model architectures, such as CNNs (high rank) and Transformers (low rank), in contrast to existing memory-efficient optimizers that applies only to a particular (rank-2) momentum tensor, e.g., linear layers. We conduct a regret bound analysis of SMMF, which shows that it converges similarly to non-memory-efficient adaptive learning rate optimizers, such as AdamNC, providing a theoretical basis for its competitive optimization capability. In our experiment, SMMF takes up to 96% less memory compared to state-of-the-art memory efficient optimizers, e.g., Adafactor, CAME, and SM3, while achieving comparable model performance on various CNN and Transformer tasks.
♻ ☆ Debiased Multimodal Understanding for Human Language Sequences AAAI2025
Human multimodal language understanding (MLU) is an indispensable component of expression analysis (e.g., sentiment or humor) from heterogeneous modalities, including visual postures, linguistic contents, and acoustic behaviours. Existing works invariably focus on designing sophisticated structures or fusion strategies to achieve impressive improvements. Unfortunately, they all suffer from the subject variation problem due to data distribution discrepancies among subjects. Concretely, MLU models are easily misled by distinct subjects with different expression customs and characteristics in the training data to learn subject-specific spurious correlations, limiting performance and generalizability across new subjects. Motivated by this observation, we introduce a recapitulative causal graph to formulate the MLU procedure and analyze the confounding effect of subjects. Then, we propose SuCI, a simple yet effective causal intervention module to disentangle the impact of subjects acting as unobserved confounders and achieve model training via true causal effects. As a plug-and-play component, SuCI can be widely applied to most methods that seek unbiased predictions. Comprehensive experiments on several MLU benchmarks clearly show the effectiveness of the proposed module.
comment: Accepted by AAAI2025
♻ ☆ LMAgent: A Large-scale Multimodal Agents Society for Multi-user Simulation
The believable simulation of multi-user behavior is crucial for understanding complex social systems. Recently, large language models (LLMs)-based AI agents have made significant progress, enabling them to achieve human-like intelligence across various tasks. However, real human societies are often dynamic and complex, involving numerous individuals engaging in multimodal interactions. In this paper, taking e-commerce scenarios as an example, we present LMAgent, a very large-scale and multimodal agents society based on multimodal LLMs. In LMAgent, besides freely chatting with friends, the agents can autonomously browse, purchase, and review products, even perform live streaming e-commerce. To simulate this complex system, we introduce a self-consistency prompting mechanism to augment agents' multimodal capabilities, resulting in significantly improved decision-making performance over the existing multi-agent system. Moreover, we propose a fast memory mechanism combined with the small-world model to enhance system efficiency, which supports more than 10,000 agent simulations in a society. Experiments on agents' behavior show that these agents achieve comparable performance to humans in behavioral indicators. Furthermore, compared with the existing LLMs-based multi-agent system, more different and valuable phenomena are exhibited, such as herd behavior, which demonstrates the potential of LMAgent in credible large-scale social behavior simulations.
♻ ☆ Radiology Report Generation via Multi-objective Preference Optimization AAAI 2025
Automatic Radiology Report Generation (RRG) is an important topic for alleviating the substantial workload of radiologists. Existing RRG approaches rely on supervised regression based on different architectures or additional knowledge injection,while the generated report may not align optimally with radiologists' preferences. Especially, since the preferences of radiologists are inherently heterogeneous and multidimensional, e.g., some may prioritize report fluency, while others emphasize clinical accuracy. To address this problem,we propose a new RRG method via Multi-objective Preference Optimization (MPO) to align the pre-trained RRG model with multiple human preferences, which can be formulated by multi-dimensional reward functions and optimized by multi-objective reinforcement learning (RL). Specifically, we use a preference vector to represent the weight of preferences and use it as a condition for the RRG model. Then, a linearly weighed reward is obtained via a dot product between the preference vector and multi-dimensional reward. Next,the RRG model is optimized to align with the preference vector by optimizing such a reward via RL. In the training stage,we randomly sample diverse preference vectors from the preference space and align the model by optimizing the weighted multi-objective rewards, which leads to an optimal policy on the entire preference space. When inference,our model can generate reports aligned with specific preferences without further fine-tuning. Extensive experiments on two public datasets show the proposed method can generate reports that cater to different preferences in a single model and achieve state-of-the-art performance.
comment: Accepted by AAAI 2025
♻ ☆ Embedding-Informed Adaptive Retrieval-Augmented Generation of Large Language Models
Retrieval-augmented large language models (LLMs) have been remarkably competent in various NLP tasks. However, it was observed by previous works that retrieval is not always helpful, especially when the LLM is already knowledgeable on the query to answer. Motivated by this, Adaptive Retrieval-Augmented Generation (ARAG) studies retrieving only when the knowledge asked by the query is absent in the LLM. Previous works of ARAG either require accessing the pre-training corpus or prompting with additional model inferences. Aiming to avoid such drawbacks, we propose to determine whether the model is knowledgeable on a query via inspecting the (contextualized) pre-trained token embeddings of LLMs. We hypothesize that such embeddings capture rich information on the model's intrinsic knowledge base, which enables an efficient way of judging the necessity to retrieve from an external corpus. Extensive experiments demonstrate our ARAG approach's superior performance across various benchmarks.
♻ ☆ Trusted Unified Feature-Neighborhood Dynamics for Multi-View Classification AAAI 2025
Multi-view classification (MVC) faces inherent challenges due to domain gaps and inconsistencies across different views, often resulting in uncertainties during the fusion process. While Evidential Deep Learning (EDL) has been effective in addressing view uncertainty, existing methods predominantly rely on the Dempster-Shafer combination rule, which is sensitive to conflicting evidence and often neglects the critical role of neighborhood structures within multi-view data. To address these limitations, we propose a Trusted Unified Feature-NEighborhood Dynamics (TUNED) model for robust MVC. This method effectively integrates local and global feature-neighborhood (F-N) structures for robust decision-making. Specifically, we begin by extracting local F-N structures within each view. To further mitigate potential uncertainties and conflicts in multi-view fusion, we employ a selective Markov random field that adaptively manages cross-view neighborhood dependencies. Additionally, we employ a shared parameterized evidence extractor that learns global consensus conditioned on local F-N structures, thereby enhancing the global integration of multi-view features. Experiments on benchmark datasets show that our method improves accuracy and robustness over existing approaches, particularly in scenarios with high uncertainty and conflicting views. The code will be made available at https://github.com/JethroJames/TUNED.
comment: Accepted to AAAI 2025
♻ ☆ Vision Language Models See What You Want but not What You See
Knowing others' intentions and taking others' perspectives are two core components of human intelligence that are typically considered to be instantiations of theory-of-mind. Infiltrating machines with these abilities is an important step towards building human-level artificial intelligence. Recently, Li et al. built CogDevelop2K, a data-intensive cognitive experiment benchmark to assess the developmental trajectory of machine intelligence. Here, to investigate intentionality understanding and perspective-taking in Vision Language Models, we leverage the IntentBench and PerspectBench of CogDevelop2K, which contains over 300 cognitive experiments grounded in real-world scenarios and classic cognitive tasks, respectively. Surprisingly, we find VLMs achieving high performance on intentionality understanding but lower performance on perspective-taking. This challenges the common belief in cognitive science literature that perspective-taking at the corresponding modality is necessary for intentionality understanding. For website see https://growing-ai-like-a-child.github.io/pages/Three%20Mountain%20Task/
comment: $\href{https://growing-ai-like-a-child.github.io/pages/Three%20Mountain%20Task/}{Website}$
♻ ☆ Leveraging Quantum Superposition to Infer the Dynamic Behavior of a Spatial-Temporal Neural Network Signaling Model
The exploration of new problem classes for quantum computation is an active area of research. In this paper, we introduce and solve a novel problem class related to dynamics on large-scale networks relevant to neurobiology and machine learning. Specifically, we ask if a network can sustain inherent dynamic activity beyond some arbitrary observation time or if the activity ceases through quiescence or saturation via an 'epileptic'-like state. We show that this class of problems can be formulated and structured to take advantage of quantum superposition and solved efficiently using the Deutsch-Jozsa and Grover quantum algorithms. To do so, we extend their functionality to address the unique requirements of how input (sub)sets into the algorithms must be mathematically structured while simultaneously constructing the inputs so that measurement outputs can be interpreted as meaningful properties of the network dynamics. This, in turn, allows us to answer the question we pose.
♻ ☆ Leveraging Chemistry Foundation Models to Facilitate Structure Focused Retrieval Augmented Generation in Multi-Agent Workflows for Catalyst and Materials Design
Molecular property prediction and generative design via deep learning models has been the subject of intense research given its potential to accelerate development of new, high-performance materials. More recently, these workflows have been significantly augmented with the advent of large language models (LLMs) and systems of autonomous agents capable of utilizing pre-trained models to make predictions in the context of more complex research tasks. While effective, there is still room for substantial improvement within agentic systems on the retrieval of salient information for material design tasks. Within this context, alternative uses of predictive deep learning models, such as leveraging their latent representations to facilitate cross-modal retrieval augmented generation within agentic systems for task-specific materials design, has remained unexplored. Herein, we demonstrate that large, pre-trained chemistry foundation models can serve as a basis for enabling structure-focused, semantic chemistry information retrieval for both small-molecules, complex polymeric materials, and reactions. Additionally, we show the use of chemistry foundation models in conjunction with multi-modal models such as OpenCLIP facilitate unprecedented queries and information retrieval across multiple characterization data domains. Finally, we demonstrate the integration of these models within multi-agent systems to facilitate structure and topological-based natural language queries and information retrieval for different research tasks.
♻ ☆ WormKAN: Are KAN Effective for Identifying and Tracking Concept Drift in Time Series?
Dynamic concepts in time series are crucial for understanding complex systems such as financial markets, healthcare, and online activity logs. These concepts help reveal structures and behaviors in sequential data for better decision-making and forecasting. However, existing models often struggle to detect and track concept drift due to limitations in interpretability and adaptability. To address this challenge, inspired by the flexibility of the recent Kolmogorov-Arnold Network (KAN), we propose WormKAN, a concept-aware KAN-based model to address concept drift in co-evolving time series. WormKAN consists of three key components: Patch Normalization, Temporal Representation Module, and Concept Dynamics. Patch normalization processes co-evolving time series into patches, treating them as fundamental modeling units to capture local dependencies while ensuring consistent scaling. The temporal representation module learns robust latent representations by leveraging a KAN-based autoencoder, complemented by a smoothness constraint, to uncover inter-patch correlations. Concept dynamics identifies and tracks dynamic transitions, revealing structural shifts in the time series through concept identification and drift detection. These transitions, akin to passing through a \textit{wormhole}, are identified by abrupt changes in the latent space. Experiments show that KAN and KAN-based models (WormKAN) effectively segment time series into meaningful concepts, enhancing the identification and tracking of concept drift.
Computation and Language 60
☆ Memory Layers at Scale
Memory layers use a trainable key-value lookup mechanism to add extra parameters to a model without increasing FLOPs. Conceptually, sparsely activated memory layers complement compute-heavy dense feed-forward layers, providing dedicated capacity to store and retrieve information cheaply. This work takes memory layers beyond proof-of-concept, proving their utility at contemporary scale. On downstream tasks, language models augmented with our improved memory layer outperform dense models with more than twice the computation budget, as well as mixture-of-expert models when matched for both compute and parameters. We find gains are especially pronounced for factual tasks. We provide a fully parallelizable memory layer implementation, demonstrating scaling laws with up to 128B memory parameters, pretrained to 1 trillion tokens, comparing to base models with up to 8B parameters.
☆ GReaTer: Gradients over Reasoning Makes Smaller Language Models Strong Prompt Optimizers
The effectiveness of large language models (LLMs) is closely tied to the design of prompts, making prompt optimization essential for enhancing their performance across a wide range of tasks. Many existing approaches to automating prompt engineering rely exclusively on textual feedback, refining prompts based solely on inference errors identified by large, computationally expensive LLMs. Unfortunately, smaller models struggle to generate high-quality feedback, resulting in complete dependence on large LLM judgment. Moreover, these methods fail to leverage more direct and finer-grained information, such as gradients, due to operating purely in text space. To this end, we introduce GReaTer, a novel prompt optimization technique that directly incorporates gradient information over task-specific reasoning. By utilizing task loss gradients, GReaTer enables self-optimization of prompts for open-source, lightweight language models without the need for costly closed-source LLMs. This allows high-performance prompt optimization without dependence on massive LLMs, closing the gap between smaller models and the sophisticated reasoning often needed for prompt refinement. Extensive evaluations across diverse reasoning tasks including BBH, GSM8k, and FOLIO demonstrate that GReaTer consistently outperforms previous state-of-the-art prompt optimization methods, even those reliant on powerful LLMs. Additionally, GReaTer-optimized prompts frequently exhibit better transferability and, in some cases, boost task performance to levels comparable to or surpassing those achieved by larger language models, highlighting the effectiveness of prompt optimization guided by gradients over reasoning. Code of GReaTer is available at https://github.com/psunlpgroup/GreaTer.
comment: 32 pages, 8 figures
☆ Human vs. AI: A Novel Benchmark and a Comparative Study on the Detection of Generated Images and the Impact of Prompts COLING 2025
With the advent of publicly available AI-based text-to-image systems, the process of creating photorealistic but fully synthetic images has been largely democratized. This can pose a threat to the public through a simplified spread of disinformation. Machine detectors and human media expertise can help to differentiate between AI-generated (fake) and real images and counteract this danger. Although AI generation models are highly prompt-dependent, the impact of the prompt on the fake detection performance has rarely been investigated yet. This work therefore examines the influence of the prompt's level of detail on the detectability of fake images, both with an AI detector and in a user study. For this purpose, we create a novel dataset, COCOXGEN, which consists of real photos from the COCO dataset as well as images generated with SDXL and Fooocus using prompts of two standardized lengths. Our user study with 200 participants shows that images generated with longer, more detailed prompts are detected significantly more easily than those generated with short prompts. Similarly, an AI-based detection model achieves better performance on images generated with longer prompts. However, humans and AI models seem to pay attention to different details, as we show in a heat map analysis.
comment: Accepted at Workshop on Detecting AI Generated Content (at COLING 2025)
☆ Formal Languages and TQFTs with Defects
A construction that assigns a Boolean 1D TQFT with defects to a finite state automaton was recently developed by Gustafson, Im, Kaldawy, Khovanov, and Lihn. We show that the construction is functorial with respect to the category of finite state automata with transducers as morphisms. Certain classes of subregular languages correspond to additional cohomological structures on the associated TQFTs. We also show that the construction generalizes to context-free grammars through a categorical version of the Chomsky-Sch\"utzenberger representation theorem, due to Melli\`es and Zeilberger. The corresponding TQFTs are then described as morphisms of colored operads on an operad of cobordisms with defects.
comment: 28 pages, 9 figures
☆ Context Canvas: Enhancing Text-to-Image Diffusion Models with Knowledge Graph-Based RAG
We introduce a novel approach to enhance the capabilities of text-to-image models by incorporating a graph-based RAG. Our system dynamically retrieves detailed character information and relational data from the knowledge graph, enabling the generation of visually accurate and contextually rich images. This capability significantly improves upon the limitations of existing T2I models, which often struggle with the accurate depiction of complex or culturally specific subjects due to dataset constraints. Furthermore, we propose a novel self-correcting mechanism for text-to-image models to ensure consistency and fidelity in visual outputs, leveraging the rich context from the graph to guide corrections. Our qualitative and quantitative experiments demonstrate that Context Canvas significantly enhances the capabilities of popular models such as Flux, Stable Diffusion, and DALL-E, and improves the functionality of ControlNet for fine-grained image editing tasks. To our knowledge, Context Canvas represents the first application of graph-based RAG in enhancing T2I models, representing a significant advancement for producing high-fidelity, context-aware multi-faceted images.
comment: Project Page: https://context-canvas.github.io/
☆ AgentTrek: Agent Trajectory Synthesis via Guiding Replay with Web Tutorials
Graphical User Interface (GUI) agents hold great potential for automating complex tasks across diverse digital environments, from web applications to desktop software. However, the development of such agents is hindered by the lack of high-quality, multi-step trajectory data required for effective training. Existing approaches rely on expensive and labor-intensive human annotation, making them unsustainable at scale. To address this challenge, we propose AgentTrek, a scalable data synthesis pipeline that generates high-quality GUI agent trajectories by leveraging web tutorials. Our method automatically gathers tutorial-like texts from the internet, transforms them into task goals with step-by-step instructions, and employs a visual-language model agent to simulate their execution in a real digital environment. A VLM-based evaluator ensures the correctness of the generated trajectories. We demonstrate that training GUI agents with these synthesized trajectories significantly improves their grounding and planning performance over the current models. Moreover, our approach is more cost-efficient compared to traditional human annotation methods. This work underscores the potential of guided replay with web tutorials as a viable strategy for large-scale GUI agent training, paving the way for more capable and autonomous digital agents.
comment: https://agenttrek.github.io
☆ TimeRefine: Temporal Grounding with Time Refining Video LLM
Video temporal grounding aims to localize relevant temporal boundaries in a video given a textual prompt. Recent work has focused on enabling Video LLMs to perform video temporal grounding via next-token prediction of temporal timestamps. However, accurately localizing timestamps in videos remains challenging for Video LLMs when relying solely on temporal token prediction. Our proposed TimeRefine addresses this challenge in two ways. First, instead of directly predicting the start and end timestamps, we reformulate the temporal grounding task as a temporal refining task: the model first makes rough predictions and then refines them by predicting offsets to the target segment. This refining process is repeated multiple times, through which the model progressively self-improves its temporal localization accuracy. Second, to enhance the model's temporal perception capabilities, we incorporate an auxiliary prediction head that penalizes the model more if a predicted segment deviates further from the ground truth, thus encouraging the model to make closer and more accurate predictions. Our plug-and-play method can be integrated into most LLM-based temporal grounding approaches. The experimental results demonstrate that TimeRefine achieves 3.6% and 5.0% mIoU improvements on the ActivityNet and Charades-STA datasets, respectively. Code and pretrained models will be released.
☆ InternLM-XComposer2.5-OmniLive: A Comprehensive Multimodal System for Long-term Streaming Video and Audio Interactions
Creating AI systems that can interact with environments over long periods, similar to human cognition, has been a longstanding research goal. Recent advancements in multimodal large language models (MLLMs) have made significant strides in open-world understanding. However, the challenge of continuous and simultaneous streaming perception, memory, and reasoning remains largely unexplored. Current MLLMs are constrained by their sequence-to-sequence architecture, which limits their ability to process inputs and generate responses simultaneously, akin to being unable to think while perceiving. Furthermore, relying on long contexts to store historical data is impractical for long-term interactions, as retaining all information becomes costly and inefficient. Therefore, rather than relying on a single foundation model to perform all functions, this project draws inspiration from the concept of the Specialized Generalist AI and introduces disentangled streaming perception, reasoning, and memory mechanisms, enabling real-time interaction with streaming video and audio input. The proposed framework InternLM-XComposer2.5-OmniLive (IXC2.5-OL) consists of three key modules: (1) Streaming Perception Module: Processes multimodal information in real-time, storing key details in memory and triggering reasoning in response to user queries. (2) Multi-modal Long Memory Module: Integrates short-term and long-term memory, compressing short-term memories into long-term ones for efficient retrieval and improved accuracy. (3) Reasoning Module: Responds to queries and executes reasoning tasks, coordinating with the perception and memory modules. This project simulates human-like cognition, enabling multimodal large language models to provide continuous and adaptive service over time.
comment: Github Repo: https://github.com/InternLM/InternLM-XComposer/tree/main/InternLM-XComposer-2.5-OmniLive
☆ OpenNER 1.0: Standardized Open-Access Named Entity Recognition Datasets in 50+ Languages
We present OpenNER 1.0, a standardized collection of openly available named entity recognition (NER) datasets. OpenNER contains 34 datasets spanning 51 languages, annotated in varying named entity ontologies. We correct annotation format issues, standardize the original datasets into a uniform representation, map entity type names to be more consistent across corpora, and provide the collection in a structure that enables research in multilingual and multi-ontology NER. We provide baseline models using three pretrained multilingual language models to compare the performance of recent models and facilitate future research in NER.
☆ DISHONEST: Dissecting misInformation Spread using Homogeneous sOcial NEtworks and Semantic Topic classification
The emergence of the COVID-19 pandemic resulted in a significant rise in the spread of misinformation on online platforms such as Twitter. Oftentimes this growth is blamed on the idea of the "echo chamber." However, the behavior said to characterize these echo chambers exists in two dimensions. The first is in a user's social interactions, where they are said to stick with the same clique of like-minded users. The second is in the content of their posts, where they are said to repeatedly espouse homogeneous ideas. In this study, we link the two by using Twitter's network of retweets to study social interactions and topic modeling to study tweet content. In order to measure the diversity of a user's interactions over time, we develop a novel metric to track the speed at which they travel through the social network. The application of these analysis methods to misinformation-focused data from the pandemic demonstrates correlation between social behavior and tweet content. We believe this correlation supports the common intuition about how antisocial users behave, and further suggests that it holds even in subcommunities already rife with misinformation.
☆ DiverseAgentEntropy: Quantifying Black-Box LLM Uncertainty through Diverse Perspectives and Multi-Agent Interaction
Quantifying the uncertainty in the factual parametric knowledge of Large Language Models (LLMs), especially in a black-box setting, poses a significant challenge. Existing methods, which gauge a model's uncertainty through evaluating self-consistency in responses to the original query, do not always capture true uncertainty. Models might respond consistently to the origin query with a wrong answer, yet respond correctly to varied questions from different perspectives about the same query, and vice versa. In this paper, we propose a novel method, DiverseAgentEntropy, for evaluating a model's uncertainty using multi-agent interaction under the assumption that if a model is certain, it should consistently recall the answer to the original query across a diverse collection of questions about the same original query. We further implement an abstention policy to withhold responses when uncertainty is high. Our method offers a more accurate prediction of the model's reliability and further detects hallucinations, outperforming other self-consistency-based methods. Additionally, it demonstrates that existing models often fail to consistently retrieve the correct answer to the same query under diverse varied questions even when knowing the correct answer.
☆ JuStRank: Benchmarking LLM Judges for System Ranking
Given the rapid progress of generative AI, there is a pressing need to systematically compare and choose between the numerous models and configurations available. The scale and versatility of such evaluations make the use of LLM-based judges a compelling solution for this challenge. Crucially, this approach requires first to validate the quality of the LLM judge itself. Previous work has focused on instance-based assessment of LLM judges, where a judge is evaluated over a set of responses, or response pairs, while being agnostic to their source systems. We argue that this setting overlooks critical factors affecting system-level ranking, such as a judge's positive or negative bias towards certain systems. To address this gap, we conduct the first large-scale study of LLM judges as system rankers. System scores are generated by aggregating judgment scores over multiple system outputs, and the judge's quality is assessed by comparing the resulting system ranking to a human-based ranking. Beyond overall judge assessment, our analysis provides a fine-grained characterization of judge behavior, including their decisiveness and bias.
Does Representation Matter? Exploring Intermediate Layers in Large Language Models
Understanding what defines a good representation in large language models (LLMs) is fundamental to both theoretical understanding and practical applications. In this paper, we investigate the quality of intermediate representations in various LLM architectures, including Transformers and State Space Models (SSMs). We find that intermediate layers often yield more informative representations for downstream tasks than the final layers. To measure the representation quality, we adapt and apply a suite of metrics - such as prompt entropy, curvature, and augmentation-invariance - originally proposed in other contexts. Our empirical study reveals significant architectural differences, how representations evolve throughout training, and how factors like input randomness and prompt length affect each layer. Notably, we observe a bimodal pattern in the entropy of some intermediate layers and consider potential explanations tied to training data. Overall, our results illuminate the internal mechanics of LLMs and guide strategies for architectural optimization and training.
comment: Accepted to 2024 NeurIPs Workshop on Machine Learning and Compression
☆ Foundational Large Language Models for Materials Research
Materials discovery and development are critical for addressing global challenges. Yet, the exponential growth in materials science literature comprising vast amounts of textual data has created significant bottlenecks in knowledge extraction, synthesis, and scientific reasoning. Large Language Models (LLMs) offer unprecedented opportunities to accelerate materials research through automated analysis and prediction. Still, their effective deployment requires domain-specific adaptation for understanding and solving domain-relevant tasks. Here, we present LLaMat, a family of foundational models for materials science developed through continued pretraining of LLaMA models on an extensive corpus of materials literature and crystallographic data. Through systematic evaluation, we demonstrate that LLaMat excels in materials-specific NLP and structured information extraction while maintaining general linguistic capabilities. The specialized LLaMat-CIF variant demonstrates unprecedented capabilities in crystal structure generation, predicting stable crystals with high coverage across the periodic table. Intriguingly, despite LLaMA-3's superior performance in comparison to LLaMA-2, we observe that LLaMat-2 demonstrates unexpectedly enhanced domain-specific performance across diverse materials science tasks, including structured information extraction from text and tables, more particularly in crystal structure generation, a potential adaptation rigidity in overtrained LLMs. Altogether, the present work demonstrates the effectiveness of domain adaptation towards developing practically deployable LLM copilots for materials research. Beyond materials science, our findings reveal important considerations for domain adaptation of LLMs, such as model selection, training methodology, and domain-specific performance, which may influence the development of specialized scientific AI systems.
☆ Systematic Analysis of LLM Contributions to Planning: Solver, Verifier, Heuristic
In this work, we provide a systematic analysis of how large language models (LLMs) contribute to solving planning problems. In particular, we examine how LLMs perform when they are used as problem solver, solution verifier, and heuristic guidance to improve intermediate solutions. Our analysis reveals that although it is difficult for LLMs to generate correct plans out-of-the-box, LLMs are much better at providing feedback signals to intermediate/incomplete solutions in the form of comparative heuristic functions. This evaluation framework provides insights into how future work may design better LLM-based tree-search algorithms to solve diverse planning and reasoning problems. We also propose a novel benchmark to evaluate LLM's ability to learn user preferences on the fly, which has wide applications in practical settings.
☆ Audios Don't Lie: Multi-Frequency Channel Attention Mechanism for Audio Deepfake Detection
With the rapid development of artificial intelligence technology, the application of deepfake technology in the audio field has gradually increased, resulting in a wide range of security risks. Especially in the financial and social security fields, the misuse of deepfake audios has raised serious concerns. To address this challenge, this study proposes an audio deepfake detection method based on multi-frequency channel attention mechanism (MFCA) and 2D discrete cosine transform (DCT). By processing the audio signal into a melspectrogram, using MobileNet V2 to extract deep features, and combining it with the MFCA module to weight different frequency channels in the audio signal, this method can effectively capture the fine-grained frequency domain features in the audio signal and enhance the Classification capability of fake audios. Experimental results show that compared with traditional methods, the model proposed in this study shows significant advantages in accuracy, precision,recall, F1 score and other indicators. Especially in complex audio scenarios, this method shows stronger robustness and generalization capabilities and provides a new idea for audio deepfake detection and has important practical application value. In the future, more advanced audio detection technologies and optimization strategies will be explored to further improve the accuracy and generalization capabilities of audio deepfake detection.
☆ The Impact of Copyrighted Material on Large Language Models: A Norwegian Perspective
The use of copyrighted materials in training generative language models raises critical legal and ethical questions. This paper presents a framework for and the results of empirically assessing the impact of copyrighted materials on the performance of large language models (LLMs) for Norwegian. We found that both books and newspapers contribute positively when the models are evaluated on a diverse set of Norwegian benchmarks, while fiction works possibly lead to decreased performance. Our experiments could inform the creation of a compensation scheme for authors whose works contribute to AI development.
comment: pre-print, under review
☆ From Intention To Implementation: Automating Biomedical Research via LLMs
Conventional biomedical research is increasingly labor-intensive due to the exponential growth of scientific literature and datasets. Artificial intelligence (AI), particularly Large Language Models (LLMs), has the potential to revolutionize this process by automating various steps. Still, significant challenges remain, including the need for multidisciplinary expertise, logicality of experimental design, and performance measurements. This paper introduces BioResearcher, the first end-to-end automated system designed to streamline the entire biomedical research process involving dry lab experiments. BioResearcher employs a modular multi-agent architecture, integrating specialized agents for search, literature processing, experimental design, and programming. By decomposing complex tasks into logically related sub-tasks and utilizing a hierarchical learning approach, BioResearcher effectively addresses the challenges of multidisciplinary requirements and logical complexity. Furthermore, BioResearcher incorporates an LLM-based reviewer for in-process quality control and introduces novel evaluation metrics to assess the quality and automation of experimental protocols. BioResearcher successfully achieves an average execution success rate of 63.07% across eight previously unmet research objectives. The generated protocols averagely outperform typical agent systems by 22.0% on five quality metrics. The system demonstrates significant potential to reduce researchers' workloads and accelerate biomedical discoveries, paving the way for future innovations in automated research systems.
☆ Unifying AI Tutor Evaluation: An Evaluation Taxonomy for Pedagogical Ability Assessment of LLM-Powered AI Tutors
In this paper, we investigate whether current state-of-the-art large language models (LLMs) are effective as AI tutors and whether they demonstrate pedagogical abilities necessary for good AI tutoring in educational dialogues. Previous efforts towards evaluation have been limited to subjective protocols and benchmarks. To bridge this gap, we propose a unified evaluation taxonomy with eight pedagogical dimensions based on key learning sciences principles, which is designed to assess the pedagogical value of LLM-powered AI tutor responses grounded in student mistakes or confusion in the mathematical domain. We release MRBench -- a new evaluation benchmark containing 192 conversations and 1,596 responses from seven state-of-the-art LLM-based and human tutors, providing gold annotations for eight pedagogical dimensions. We assess reliability of the popular Prometheus2 LLM as an evaluator and analyze each tutor's pedagogical abilities, highlighting which LLMs are good tutors and which ones are more suitable as question-answering systems. We believe that the presented taxonomy, benchmark, and human-annotated labels will streamline the evaluation process and help track the progress in AI tutors' development.
comment: 8 pages
☆ Text Generation Models for Luxembourgish with Limited Data: A Balanced Multilingual Strategy
This paper addresses the challenges in developing language models for less-represented languages, with a focus on Luxembourgish. Despite its active development, Luxembourgish faces a digital data scarcity, exacerbated by Luxembourg's multilingual context. We propose a novel text generation model based on the T5 architecture, combining limited Luxembourgish data with equal amounts, in terms of size and type, of German and French data. We hypothesise that a model trained on Luxembourgish, German, and French will improve the model's cross-lingual transfer learning capabilities and outperform monolingual and large multilingual models. To verify this, the study at hand explores whether multilingual or monolingual training is more beneficial for Luxembourgish language generation. For the evaluation, we introduce LuxGen, a text generation benchmark that is the first of its kind for Luxembourgish.
comment: Accepted at VarDial 2025
☆ Imitate, Explore, and Self-Improve: A Reproduction Report on Slow-thinking Reasoning Systems
Recently, slow-thinking reasoning systems, such as o1, have demonstrated remarkable capabilities in solving complex reasoning tasks. These systems typically engage in an extended thinking process before responding to a query, allowing them to generate more thorough, accurate, and well-reasoned solutions. These systems are primarily developed and maintained by industry, with their core techniques not publicly disclosed. In response, an increasing number of studies from the research community aim to explore the technical foundations underlying these powerful reasoning systems. Building on these prior efforts, this paper presents a reproduction report on implementing o1-like reasoning systems. We introduce an "imitate, explore, and self-improve" framework as our primary technical approach to train the reasoning model. In the initial phase, we use distilled long-form thought data to fine-tune the reasoning model, enabling it to invoke a slow-thinking mode. The model is then encouraged to explore challenging problems by generating multiple rollouts, which can result in increasingly more high-quality trajectories that lead to correct answers. Furthermore, the model undergoes self-improvement by iteratively refining its training dataset. To verify the effectiveness of this approach, we conduct extensive experiments on three challenging benchmarks. The experimental results demonstrate that our approach achieves competitive performance compared to industry-level reasoning systems on these benchmarks.
comment: Technical Report on Slow Thinking with LLMs: Part II
☆ From Bench to Bedside: A Review of Clinical Trialsin Drug Discovery and Development
Clinical trials are an indispensable part of the drug development process, bridging the gap between basic research and clinical application. During the development of new drugs, clinical trials are used not only to evaluate the safety and efficacy of the drug but also to explore its dosage, treatment regimens, and potential side effects. This review discusses the various stages of clinical trials, including Phase I (safety assessment), Phase II (preliminary efficacy evaluation), Phase III (large-scale validation), and Phase IV (post-marketing surveillance), highlighting the characteristics of each phase and their interrelationships. Additionally, the paper addresses the major challenges encountered in clinical trials, such as ethical issues, subject recruitment difficulties, diversity and representativeness concerns, and proposes strategies for overcoming these challenges. With the advancement of technology, innovative technologies such as artificial intelligence, big data, and digitalization are gradually transforming clinical trial design and implementation, improving trial efficiency and data quality. The article also looks forward to the future of clinical trials, particularly the impact of emerging therapies such as gene therapy and immunotherapy on trial design, as well as the importance of regulatory reforms and global collaboration. In conclusion, the core role of clinical trials in drug development will continue to drive the progress of innovative drug development and clinical treatment.
comment: 11 pages
☆ Word Sense Linking: Disambiguating Outside the Sandbox
Word Sense Disambiguation (WSD) is the task of associating a word in a given context with its most suitable meaning among a set of possible candidates. While the task has recently witnessed renewed interest, with systems achieving performances above the estimated inter-annotator agreement, at the time of writing it still struggles to find downstream applications. We argue that one of the reasons behind this is the difficulty of applying WSD to plain text. Indeed, in the standard formulation, models work under the assumptions that a) all the spans to disambiguate have already been identified, and b) all the possible candidate senses of each span are provided, both of which are requirements that are far from trivial. In this work, we present a new task called Word Sense Linking (WSL) where, given an input text and a reference sense inventory, systems have to both identify which spans to disambiguate and then link them to their most suitable meaning.We put forward a transformer-based architecture for the task and thoroughly evaluate both its performance and those of state-of-the-art WSD systems scaled to WSL, iteratively relaxing the assumptions of WSD. We hope that our work will foster easier integration of lexical semantics into downstream applications.
☆ Falcon-UI: Understanding GUI Before Following User Instructions
Pursuing human-like interaction for Graphical User Interface (GUI) agents requires understanding the GUI context and following user instructions. However, existing works typically couple these two aspects and focus more on instruct-following abilities, while ignoring the importance of understanding the GUI context. In this paper, we introduce an instruction-free GUI navigation dataset, termed Insight-UI Dataset, to enhance model comprehension of GUI environments. Insight-UI Dataset is automatically generated from the Common Crawl corpus, simulating various platforms -- including iOS, Android, Windows, and Linux -- across multiple resolutions on 312K domains. Although GUI interactions vary by context, diverse interfaces share common internal patterns, such as clicking an item to view its details. It implies the feasibility of independent GUI operation learning, followed by joint optimization with instruction tuning. Thereby, we develop the GUI agent model Falcon-UI, which is initially pretrained on Insight-UI Dataset and subsequently fine-tuned on Android and Web GUI datasets, including AITW, AITZ, Android Control, and Mind2Web. With 7 billion parameters, Falcon-UI achieves accuracy comparable to the 72 billion-parameter Qwen2VL on AITZ, validating the alignment between GUI context comprehension and agent performance. Our code and dataset will be open-sourced.
comment: 18 pages, 14 figures
☆ Causal Graphical Models for Vision-Language Compositional Understanding
Recent work has empirically shown that Vision-Language Models (VLMs) struggle to fully understand the compositional properties of the human language, usually modeling an image caption as a "bag of words". As a result, they perform poorly on compositional tasks, which require a deeper understanding of the different entities of a sentence (subject, verb, etc.) jointly with their mutual relationships in order to be solved. In this paper, we model the dependency relations among textual and visual tokens using a Causal Graphical Model (CGM), built using a dependency parser, and we train a decoder conditioned by the VLM visual encoder. Differently from standard autoregressive or parallel predictions, our decoder's generative process is partially-ordered following the CGM structure. This structure encourages the decoder to learn only the main causal dependencies in a sentence discarding spurious correlations. Using extensive experiments on five compositional benchmarks, we show that our method significantly outperforms all the state-of-the-art compositional approaches by a large margin, and it also improves over methods trained using much larger datasets.
☆ Training LayoutLM from Scratch for Efficient Named-Entity Recognition in the Insurance Domain
Generic pre-trained neural networks may struggle to produce good results in specialized domains like finance and insurance. This is due to a domain mismatch between training data and downstream tasks, as in-domain data are often scarce due to privacy constraints. In this work, we compare different pre-training strategies for LayoutLM. We show that using domain-relevant documents improves results on a named-entity recognition (NER) problem using a novel dataset of anonymized insurance-related financial documents called Payslips. Moreover, we show that we can achieve competitive results using a smaller and faster model.
comment: Coling 2025 workshop (FinNLP)
☆ CRVQ: Channel-relaxed Vector Quantization for Extreme Compression of LLMs
Powerful large language models (LLMs) are increasingly expected to be deployed with lower computational costs, enabling their capabilities on resource-constrained devices. Post-training quantization (PTQ) has emerged as a star approach to achieve this ambition, with best methods compressing weights to less than 2 bit on average. In this paper, we propose Channel-Relaxed Vector Quantization (CRVQ), a novel technique that significantly improves the performance of PTQ baselines at the cost of only minimal additional bits. This state-of-the-art extreme compression method achieves its results through two key innovations: (1) carefully selecting and reordering a very small subset of critical weight channels, and (2) leveraging multiple codebooks to relax the constraint of critical channels. With our method, we demonstrate a 38.9% improvement over the current strongest sub-2-bit PTQ baseline, enabling nearer lossless 1-bit compression. Furthermore, our approach offers flexible customization of quantization bit-width and performance, providing a wider range of deployment options for diverse hardware platforms.
comment: 5 figures, 4 tables
☆ Learning to Solve Domain-Specific Calculation Problems with Knowledge-Intensive Programs Generator
Domain Large Language Models (LLMs) are developed for domain-specific tasks based on general LLMs. But it still requires professional knowledge to facilitate the expertise for some domain-specific tasks. In this paper, we investigate into knowledge-intensive calculation problems. We find that the math problems to be challenging for LLMs, when involving complex domain-specific rules and knowledge documents, rather than simple formulations of terminologies. Therefore, we propose a pipeline to solve the domain-specific calculation problems with Knowledge-Intensive Programs Generator more effectively, named as KIPG. It generates knowledge-intensive programs according to the domain-specific documents. For each query, key variables are extracted, then outcomes which are dependent on domain knowledge are calculated with the programs. By iterative preference alignment, the code generator learns to improve the logic consistency with the domain knowledge. Taking legal domain as an example, we have conducted experiments to prove the effectiveness of our pipeline, and extensive analysis on the modules. We also find that the code generator is also adaptable to other domains, without training on the new knowledge.
comment: Under review
☆ Towards Understanding the Robustness of LLM-based Evaluations under Perturbations
Traditional evaluation metrics like BLEU and ROUGE fall short when capturing the nuanced qualities of generated text, particularly when there is no single ground truth. In this paper, we explore the potential of Large Language Models (LLMs), specifically Google Gemini 1, to serve as automatic evaluators for non-standardized metrics in summarization and dialog-based tasks. We conduct experiments across multiple prompting strategies to examine how LLMs fare as quality evaluators when compared with human judgments on the SummEval and USR datasets, asking the model to generate both a score as well as a justification for the score. Furthermore, we explore the robustness of the LLM evaluator by using perturbed inputs. Our findings suggest that while LLMs show promise, their alignment with human evaluators is limited, they are not robust against perturbations and significant improvements are required for their standalone use as reliable evaluators for subjective metrics.
comment: Accepted at ICON 2024
☆ Make Satire Boring Again: Reducing Stylistic Bias of Satirical Corpus by Utilizing Generative LLMs COLING2025
Satire detection is essential for accurately extracting opinions from textual data and combating misinformation online. However, the lack of diverse corpora for satire leads to the problem of stylistic bias which impacts the models' detection performances. This study proposes a debiasing approach for satire detection, focusing on reducing biases in training data by utilizing generative large language models. The approach is evaluated in both cross-domain (irony detection) and cross-lingual (English) settings. Results show that the debiasing method enhances the robustness and generalizability of the models for satire and irony detection tasks in Turkish and English. However, its impact on causal language models, such as Llama-3.1, is limited. Additionally, this work curates and presents the Turkish Satirical News Dataset with detailed human annotations, with case studies on classification, debiasing, and explainability.
comment: Accepted to BUCC2025 Workshop @COLING2025
☆ CleanComedy: Creating Friendly Humor through Generative Techniques
Humor generation is a challenging task in natural language processing due to limited resources and the quality of existing datasets. Available humor language resources often suffer from toxicity and duplication, limiting their effectiveness for training robust models. This paper proposes CleanComedy, a specialized, partially annotated toxicity-filtered corpus of English and Russian jokes collected from various sources. We study the effectiveness of our data filtering approach through a survey on humor and toxicity levels in various joke groups. In addition, we study advances in computer humor generation by comparing jokes written by humans with various groups of generative jokes, including our baseline models trained on the CleanComedy datasets.
☆ ReFF: Reinforcing Format Faithfulness in Language Models across Varied Tasks AAAI 2025
Following formatting instructions to generate well-structured content is a fundamental yet often unmet capability for large language models (LLMs). To study this capability, which we refer to as format faithfulness, we present FormatBench, a comprehensive format-related benchmark. Compared to previous format-related benchmarks, FormatBench involves a greater variety of tasks in terms of application scenes (traditional NLP tasks, creative works, autonomous agency tasks), human-LLM interaction styles (single-turn instruction, multi-turn chat), and format types (inclusion, wrapping, length, coding). Moreover, each task in FormatBench is attached with a format checker program. Extensive experiments on the benchmark reveal that state-of-the-art open- and closed-source LLMs still suffer from severe deficiency in format faithfulness. By virtue of the decidable nature of formats, we propose to Reinforce Format Faithfulness (ReFF) to help LLMs generate formatted output as instructed without compromising general quality. Without any annotated data, ReFF can substantially improve the format faithfulness rate (e.g., from 21.6% in original LLaMA3 to 95.0% on caption segmentation task), while keep the general quality comparable (e.g., from 47.3 to 46.4 in F1 scores). Combined with labeled training data, ReFF can simultaneously improve both format faithfulness (e.g., from 21.6% in original LLaMA3 to 75.5%) and general quality (e.g., from 47.3 to 61.6 in F1 scores). We further offer an interpretability analysis to explain how ReFF improves both format faithfulness and general quality.
comment: Accepted to AAAI 2025
☆ When Text Embedding Meets Large Language Model: A Comprehensive Survey
Text embedding has become a foundational technology in natural language processing (NLP) during the deep learning era, driving advancements across a wide array of downstream tasks. While many natural language understanding challenges can now be modeled using generative paradigms and leverage the robust generative and comprehension capabilities of large language models (LLMs), numerous practical applications, such as semantic matching, clustering, and information retrieval, continue to rely on text embeddings for their efficiency and effectiveness. In this survey, we categorize the interplay between LLMs and text embeddings into three overarching themes: (1) LLM-augmented text embedding, enhancing traditional embedding methods with LLMs; (2) LLMs as text embedders, utilizing their innate capabilities for embedding generation; and (3) Text embedding understanding with LLMs, leveraging LLMs to analyze and interpret embeddings. By organizing these efforts based on interaction patterns rather than specific downstream applications, we offer a novel and systematic overview of contributions from various research and application domains in the era of LLMs. Furthermore, we highlight the unresolved challenges that persisted in the pre-LLM era with pre-trained language models (PLMs) and explore the emerging obstacles brought forth by LLMs. Building on this analysis, we outline prospective directions for the evolution of text embedding, addressing both theoretical and practical opportunities in the rapidly advancing landscape of NLP.
comment: Work in progress
☆ PolyIPA -- Multilingual Phoneme-to-Grapheme Conversion Model
This paper presents PolyIPA, a novel multilingual phoneme-to-grapheme conversion model designed for multilingual name transliteration, onomastic research, and information retrieval. The model leverages two helper models developed for data augmentation: IPA2vec for finding soundalikes across languages, and similarIPA for handling phonetic notation variations. Evaluated on a test set that spans multiple languages and writing systems, the model achieves a mean Character Error Rate of 0.055 and a character-level BLEU score of 0.914, with particularly strong performance on languages with shallow orthographies. The implementation of beam search further improves practical utility, with top-3 candidates reducing the effective error rate by 52.7\% (to CER: 0.026), demonstrating the model's effectiveness for cross-linguistic applications.
♻ ☆ Training LLMs over Neurally Compressed Text
In this paper, we explore the idea of training large language models (LLMs) over highly compressed text. While standard subword tokenizers compress text by a small factor, neural text compressors can achieve much higher rates of compression. If it were possible to train LLMs directly over neurally compressed text, this would confer advantages in training and serving efficiency, as well as easier handling of long text spans. The main obstacle to this goal is that strong compression tends to produce opaque outputs that are not well-suited for learning. In particular, we find that text na\"ively compressed via Arithmetic Coding is not readily learnable by LLMs. To overcome this, we propose Equal-Info Windows, a novel compression technique whereby text is segmented into blocks that each compress to the same bit length. Using this method, we demonstrate effective learning over neurally compressed text that improves with scale, and outperforms byte-level baselines by a wide margin on perplexity and inference speed benchmarks. While our method delivers worse perplexity than subword tokenizers for models trained with the same parameter count, it has the benefit of shorter sequence lengths. Shorter sequence lengths require fewer autoregressive generation steps, and reduce latency. Finally, we provide extensive analysis of the properties that contribute to learnability, and offer concrete suggestions for how to further improve the performance of high-compression tokenizers.
comment: Accepted in TMLR https://openreview.net/forum?id=pRvhMSV48t
♻ ☆ Test Case-Informed Knowledge Tracing for Open-ended Coding Tasks
Open-ended coding tasks, which ask students to construct programs according to certain specifications, are common in computer science education. Student modeling can be challenging since their open-ended nature means that student code can be diverse. Traditional knowledge tracing (KT) models that only analyze response correctness may not fully capture nuances in student knowledge from student code. In this paper, we introduce Test case-Informed Knowledge Tracing for Open-ended Coding (TIKTOC), a framework to simultaneously analyze and predict both open-ended student code and whether the code passes each test case. We augment the existing CodeWorkout dataset with the test cases used for a subset of the open-ended coding questions, and propose a multi-task learning KT method to simultaneously analyze and predict 1) whether a student's code submission passes each test case and 2) the student's open-ended code, using a large language model as the backbone. We quantitatively show that these methods outperform existing KT methods for coding that only use the overall score a code submission receives. We also qualitatively demonstrate how test case information, combined with open-ended code, helps us gain fine-grained insights into student knowledge.
comment: Published in LAK 2025: The 15th International Learning Analytics and Knowledge Conference
♻ ☆ Piecing It All Together: Verifying Multi-Hop Multimodal Claims COLING 2025
Existing claim verification datasets often do not require systems to perform complex reasoning or effectively interpret multimodal evidence. To address this, we introduce a new task: multi-hop multimodal claim verification. This task challenges models to reason over multiple pieces of evidence from diverse sources, including text, images, and tables, and determine whether the combined multimodal evidence supports or refutes a given claim. To study this task, we construct MMCV, a large-scale dataset comprising 15k multi-hop claims paired with multimodal evidence, generated and refined using large language models, with additional input from human feedback. We show that MMCV is challenging even for the latest state-of-the-art multimodal large language models, especially as the number of reasoning hops increases. Additionally, we establish a human performance benchmark on a subset of MMCV. We hope this dataset and its evaluation task will encourage future research in multimodal multi-hop claim verification.
comment: COLING 2025
♻ ☆ Unveiling the Impact of Coding Data Instruction Fine-Tuning on Large Language Models Reasoning
Instruction Fine-Tuning (IFT) significantly enhances the zero-shot capabilities of pretrained Large Language Models (LLMs). While coding data is known to boost LLM reasoning abilities during pretraining, its role in activating internal reasoning capacities during IFT remains understudied. This paper investigates a key question: How does coding data impact LLMs' reasoning capacities during IFT stage? To explore this, we thoroughly examine the impact of coding data across different coding data proportions, model families, sizes, and reasoning domains, from various perspectives. Specifically, we create three IFT datasets with increasing coding data proportions, fine-tune six LLM backbones across different families and scales on these datasets, evaluate the tuned models' performance across twelve tasks in three reasoning domains, and analyze the outcomes from three broad-to-granular perspectives: overall, domain-level, and task-specific. Our holistic analysis provides valuable insights into each perspective. First, coding data tuning enhances the overall reasoning capabilities of LLMs across different model families and scales. Moreover, while the impact of coding data varies by domain, it shows consistent trends within each domain across different model families and scales. Additionally, coding data generally provides comparable task-specific benefits across model families, with optimal proportions in IFT datasets being task-dependent.
♻ ☆ Altogether: Image Captioning via Re-aligning Alt-text EMNLP 2024
This paper focuses on creating synthetic data to improve the quality of image captions. Existing works typically have two shortcomings. First, they caption images from scratch, ignoring existing alt-text metadata, and second, lack transparency if the captioners' training data (e.g. GPT) is unknown. In this paper, we study a principled approach Altogether based on the key idea to edit and re-align existing alt-texts associated with the images. To generate training data, we perform human annotation where annotators start with the existing alt-text and re-align it to the image content in multiple rounds, consequently constructing captions with rich visual concepts. This differs from prior work that carries out human annotation as a one-time description task solely based on images and annotator knowledge. We train a captioner on this data that generalizes the process of re-aligning alt-texts at scale. Our results show our Altogether approach leads to richer image captions that also improve text-to-image generation and zero-shot image classification tasks.
comment: accepted by EMNLP 2024; Meta CLIP 1.2 Data Engine
♻ ☆ LCFO: Long Context and Long Form Output Dataset and Benchmarking
This paper presents the Long Context and Form Output (LCFO) benchmark, a novel evaluation framework for assessing gradual summarization and summary expansion capabilities across diverse domains. LCFO consists of long input documents (5k words average length), each of which comes with three summaries of different lengths (20%, 10%, and 5% of the input text), as well as approximately 15 questions and answers (QA) related to the input content. Notably, LCFO also provides alignments between specific QA pairs and corresponding summaries in 7 domains. The primary motivation behind providing summaries of different lengths is to establish a controllable framework for generating long texts from shorter inputs, i.e. summary expansion. To establish an evaluation metric framework for summarization and summary expansion, we provide human evaluation scores for human-generated outputs, as well as results from various state-of-the-art large language models (LLMs). GPT-4o-mini achieves best human scores among automatic systems in both summarization and summary expansion tasks (~ +10% and +20%, respectively). It even surpasses human output quality in the case of short summaries (~ +7%). Overall automatic metrics achieve low correlations with human evaluation scores (~ 0.4) but moderate correlation on specific evaluation aspects such as fluency and attribution (~ 0.6). The LCFO benchmark offers a standardized platform for evaluating summarization and summary expansion performance, as well as corresponding automatic metrics, thereby providing an important evaluation framework to advance generative AI.
♻ ☆ Improving the Validity of Automatically Generated Feedback via Reinforcement Learning
Automatically generating feedback via large language models (LLMs) in intelligent tutoring systems and online learning platforms has the potential to improve the learning outcomes of many students. However, both feedback generation and evaluation are challenging: feedback content has to be valid especially in subjects like math, which requires models to understand the problem, the solution, and where the student's error lies. Feedback also has to be pedagogically valid to reflect effective tutoring strategies, such as explaining possible misconceptions and encouraging the student, among other desirable features. In this work, we address both problems of automatically generating and evaluating feedback while considering both correctness and alignment. First, we propose a rubric for evaluating math feedback and show that GPT-4 is able to effectively use it to annotate human-written and LLM-generated feedback. Second, we propose a framework for feedback generation that optimizes both correctness and alignment using reinforcement learning (RL). Specifically, we use GPT-4's annotations to create preferences over feedback pairs in an augmented dataset for training via direct preference optimization (DPO). We show that our methods significantly increase the correctness and alignment of generated feedback with Llama 2, an open-source LLM, qualitatively analyze our generation and evaluation systems using case studies, and outline several areas for future work.
comment: Best student paper award, Published in AIED 2024: The 25th International Conference on Artificial Intelligence in Education
♻ ☆ Evaluating GPT-4 at Grading Handwritten Solutions in Math Exams
Recent advances in generative artificial intelligence (AI) have shown promise in accurately grading open-ended student responses. However, few prior works have explored grading handwritten responses due to a lack of data and the challenge of combining visual and textual information. In this work, we leverage state-of-the-art multi-modal AI models, in particular GPT-4o, to automatically grade handwritten responses to college-level math exams. Using real student responses to questions in a probability theory exam, we evaluate GPT-4o's alignment with ground-truth scores from human graders using various prompting techniques. We find that while providing rubrics improves alignment, the model's overall accuracy is still too low for real-world settings, showing there is significant room for growth in this task.
comment: Published in LAK 2025: The 15th International Learning Analytics and Knowledge Conference
♻ ☆ SAGED: A Holistic Bias-Benchmarking Pipeline for Language Models with Customisable Fairness Calibration COLING 2025
The development of unbiased large language models is widely recognized as crucial, yet existing benchmarks fall short in detecting biases due to limited scope, contamination, and lack of a fairness baseline. SAGED(bias) is the first holistic benchmarking pipeline to address these problems. The pipeline encompasses five core stages: scraping materials, assembling benchmarks, generating responses, extracting numeric features, and diagnosing with disparity metrics. SAGED includes metrics for max disparity, such as impact ratio, and bias concentration, such as Max Z-scores. Noticing that metric tool bias and contextual bias in prompts can distort evaluation, SAGED implements counterfactual branching and baseline calibration for mitigation. For demonstration, we use SAGED on G20 Countries with popular 8b-level models including Gemma2, Llama3.1, Mistral, and Qwen2. With sentiment analysis, we find that while Mistral and Qwen2 show lower max disparity and higher bias concentration than Gemma2 and Llama3.1, all models are notably biased against countries like Russia and (except for Qwen2) China. With further experiments to have models role-playing U.S. presidents, we see bias amplifies and shifts in heterogeneous directions. Moreover, we see Qwen2 and Mistral not engage in role-playing, while Llama3.1 and Gemma2 role-play Trump notably more intensively than Biden and Harris, indicating role-playing performance bias in these models.
comment: COLING 2025 Main Conference
♻ ☆ Few-Shot Domain Adaptation for Named-Entity Recognition via Joint Constrained k-Means and Subspace Selection COLING 2025
Named-entity recognition (NER) is a task that typically requires large annotated datasets, which limits its applicability across domains with varying entity definitions. This paper addresses few-shot NER, aiming to transfer knowledge to new domains with minimal supervision. Unlike previous approaches that rely solely on limited annotated data, we propose a weakly supervised algorithm that combines small labeled datasets with large amounts of unlabeled data. Our method extends the k-means algorithm with label supervision, cluster size constraints and domain-specific discriminative subspace selection. This unified framework achieves state-of-the-art results in few-shot NER on several English datasets.
comment: COLING 2025
♻ ☆ Dspy-based Neural-Symbolic Pipeline to Enhance Spatial Reasoning in LLMs
Large Language Models (LLMs) have demonstrated remarkable capabilities across various tasks, yet they often struggle with spatial reasoning. This paper presents a novel neural-symbolic framework that enhances LLMs' spatial reasoning abilities through iterative feedback between LLMs and Answer Set Programming (ASP). We evaluate our approach on two benchmark datasets: StepGame and SparQA, implementing three distinct strategies: (1) direct prompting baseline, (2) Facts+Rules prompting, and (3) DSPy-based LLM+ASP pipeline with iterative refinement. Our experimental results demonstrate that the LLM+ASP pipeline significantly outperforms baseline methods, achieving an average 82% accuracy on StepGame and 69% on SparQA, marking improvements of 40-50% and 8-15% respectively over direct prompting. The success stems from three key innovations: (1) effective separation of semantic parsing and logical reasoning through a modular pipeline, (2) iterative feedback mechanism between LLMs and ASP solvers that improves program rate, and (3) robust error handling that addresses parsing, grounding, and solving failures. Additionally, we propose Facts+Rules as a lightweight alternative that achieves comparable performance on complex SparQA dataset, while reducing computational overhead.Our analysis across different LLM architectures (Deepseek, Llama3-70B, GPT-4.0 mini) demonstrates the framework's generalizability and provides insights into the trade-offs between implementation complexity and reasoning capability, contributing to the development of more interpretable and reliable AI systems.
♻ ☆ EVQAScore: Efficient Video Question Answering Data Evaluation
Video question-answering (QA) is a core task in video understanding. Evaluating the quality of video QA and video caption data quality for training video large language models (VideoLLMs) is an essential challenge. Although various methods have been proposed for assessing video caption quality, there remains a lack of dedicated evaluation methods for Video QA. To address this gap, we introduce EVQAScore, a reference-free method that leverages keyword extraction to assess both video caption and video QA data quality. Additionally, we incorporate frame sampling and rescaling techniques to enhance the efficiency and robustness of our evaluation, this enables our score to evaluate the quality of extremely long videos. Our approach achieves state-of-the-art (SOTA) performance (32.8 for Kendall correlation and 42.3 for Spearman correlation, 4.7 and 5.9 higher than the previous method PAC-S++) on the VATEX-EVAL benchmark for video caption evaluation. Furthermore, by using EVQAScore for data selection, we achieved SOTA results with only 12.5\% of the original data volume, outperforming the previous SOTA method PAC-S and 100\% of data.
♻ ☆ Detection of Non-recorded Word Senses in English and Swedish
This study addresses the task of Unknown Sense Detection in English and Swedish. The primary objective of this task is to determine whether the meaning of a particular word usage is documented in a dictionary or not. For this purpose, sense entries are compared with word usages from modern and historical corpora using a pre-trained Word-in-Context embedder that allows us to model this task in a few-shot scenario. Additionally, we use human annotations on the target corpora to adapt hyperparameters and evaluate our models using 5-fold cross-validation. Compared to a random sample from a corpus, our model is able to considerably increase the detected number of word usages with non-recorded senses.
comment: 9 pages
♻ ☆ Importance Weighting Can Help Large Language Models Self-Improve
Large language models (LLMs) have shown remarkable capability in numerous tasks and applications. However, fine-tuning LLMs using high-quality datasets under external supervision remains prohibitively expensive. In response, LLM self-improvement approaches have been vibrantly developed recently. The typical paradigm of LLM self-improvement involves training LLM on self-generated data, part of which may be detrimental and should be filtered out due to the unstable data quality. While current works primarily employs filtering strategies based on answer correctness, in this paper, we demonstrate that filtering out correct but with high distribution shift extent (DSE) samples could also benefit the results of self-improvement. Given that the actual sample distribution is usually inaccessible, we propose a new metric called DS weight to approximate DSE, inspired by the Importance Weighting methods. Consequently, we integrate DS weight with self-consistency to comprehensively filter the self-generated samples and fine-tune the language model. Experiments show that with only a tiny valid set (up to 5\% size of the training set) to compute DS weight, our approach can notably promote the reasoning ability of current LLM self-improvement methods. The resulting performance is on par with methods that rely on external supervision from pre-trained reward models.
♻ ☆ Chatbots im Schulunterricht: Wir testen das Fobizz-Tool zur automatischen Bewertung von Hausaufgaben
[Study in German language.] This study examines the AI-powered grading tool "AI Grading Assistant" by the German company Fobizz, designed to support teachers in evaluating and providing feedback on student assignments. Against the societal backdrop of an overburdened education system and rising expectations for artificial intelligence as a solution to these challenges, the investigation evaluates the tool's functional suitability through two test series. The results reveal significant shortcomings: The tool's numerical grades and qualitative feedback are often random and do not improve even when its suggestions are incorporated. The highest ratings are achievable only with texts generated by ChatGPT. False claims and nonsensical submissions frequently go undetected, while the implementation of some grading criteria is unreliable and opaque. Since these deficiencies stem from the inherent limitations of large language models (LLMs), fundamental improvements to this or similar tools are not immediately foreseeable. The study critiques the broader trend of adopting AI as a quick fix for systemic problems in education, concluding that Fobizz's marketing of the tool as an objective and time-saving solution is misleading and irresponsible. Finally, the study calls for systematic evaluation and subject-specific pedagogical scrutiny of the use of AI tools in educational contexts.
comment: 33 pages, in German language
♻ ☆ Exploring Language Model Generalization in Low-Resource Extractive QA COLING 2025
In this paper, we investigate Extractive Question Answering (EQA) with Large Language Models (LLMs) under domain drift, i.e., can LLMs generalize to domains that require specific knowledge such as medicine and law in a zero-shot fashion without additional in-domain training? To this end, we devise a series of experiments to explain the performance gap empirically. Our findings suggest that: (a) LLMs struggle with dataset demands of closed domains such as retrieving long answer spans; (b) Certain LLMs, despite showing strong overall performance, display weaknesses in meeting basic requirements as discriminating between domain-specific senses of words which we link to pre-processing decisions; (c) Scaling model parameters is not always effective for cross domain generalization; and (d) Closed-domain datasets are quantitatively much different than open-domain EQA datasets and current LLMs struggle to deal with them. Our findings point out important directions for improving existing LLMs.
comment: Accepted to COLING 2025
♻ ☆ TOP-Training: Target-Oriented Pretraining for Medical Extractive Question Answering COLING 2025
We study extractive question-answering in the medical domain (Medical-EQA). This problem has two main challenges: (i) domain specificity, as most AI models lack necessary domain knowledge, and (ii) extraction-based answering style, which restricts most autoregressive LLMs due to potential hallucinations. To handle those challenges, we propose TOP-Training, a target-oriented pre-training paradigm that stands out among all domain adaptation techniques with two desirable features: (i) TOP-Training moves one step further than popular domain-oriented fine-tuning since it not only moves closer to the target domain, but also familiarizes itself with the target dataset, and (ii) it does not assume the existence of a large set of unlabeled instances from the target domain. Specifically, for a target Medical-EQA dataset, we extract its entities and leverage large language models (LLMs) to generate synthetic texts containing those entities; we then demonstrate that pretraining on this synthetic text data yields better performance on the target Medical-EQA benchmarks. Overall, our contributions are threefold: (i) TOP-Training, a new pretraining technique to effectively adapt LLMs to better solve a target problem, (ii) TOP-Training has a wide application scope because it does not require the target problem to have a large set of unlabeled data, and (iii) our experiments highlight the limitations of autoregressive LLMs, emphasizing TOP-Training as a means to unlock the true potential of bidirectional LLMs.
comment: Accepted to COLING 2025
♻ ☆ Empathy Level Alignment via Reinforcement Learning for Empathetic Response Generation
Empathetic response generation, aiming to understand the user's situation and feelings and respond empathically, is crucial in building human-like dialogue systems. Traditional approaches typically employ maximum likelihood estimation as the optimization objective during training, yet fail to align the empathy levels between generated and target responses. To this end, we propose an empathetic response generation framework using reinforcement learning (EmpRL). The framework develops an effective empathy reward function and generates empathetic responses by maximizing the expected reward through reinforcement learning. EmpRL utilizes the pre-trained T5 model as the generator and further fine-tunes it to initialize the policy. To align the empathy levels between generated and target responses within a given context, an empathy reward function containing three empathy communication mechanisms -- emotional reaction, interpretation, and exploration -- is constructed using pre-designed and pre-trained empathy identifiers. During reinforcement learning training, the proximal policy optimization algorithm is used to fine-tune the policy, enabling the generation of empathetic responses. Both automatic and human evaluations demonstrate that the proposed EmpRL framework significantly improves the quality of generated responses, enhances the similarity in empathy levels between generated and target responses, and produces empathetic responses covering both affective and cognitive aspects.
♻ ☆ How Likely Do LLMs with CoT Mimic Human Reasoning? COLING 2025
Chain-of-thought emerges as a promising technique for eliciting reasoning capabilities from Large Language Models (LLMs). However, it does not always improve task performance or accurately represent reasoning processes, leaving unresolved questions about its usage. In this paper, we diagnose the underlying mechanism by comparing the reasoning process of LLMs with humans, using causal analysis to understand the relationships between the problem instruction, reasoning, and the answer in LLMs. Our empirical study reveals that LLMs often deviate from the ideal causal chain, resulting in spurious correlations and potential consistency errors (inconsistent reasoning and answers). We also examine various factors influencing the causal structure, finding that in-context learning with examples strengthens it, while post-training techniques like supervised fine-tuning and reinforcement learning on human feedback weaken it. To our surprise, the causal structure cannot be strengthened by enlarging the model size only, urging research on new techniques. We hope that this preliminary study will shed light on understanding and improving the reasoning process in LLM.
comment: COLING 2025 Camera Version (8 pages, 3 figures, 18 tables)
♻ ☆ NLPineers@ NLU of Devanagari Script Languages 2025: Hate Speech Detection using Ensembling of BERT-based models
This paper explores hate speech detection in Devanagari-scripted languages, focusing on Hindi and Nepali, for Subtask B of the CHIPSAL@COLING 2025 Shared Task. Using a range of transformer-based models such as XLM-RoBERTa, MURIL, and IndicBERT, we examine their effectiveness in navigating the nuanced boundary between hate speech and free expression. Our best performing model, implemented as ensemble of multilingual BERT models achieve Recall of 0.7762 (Rank 3/31 in terms of recall) and F1 score of 0.6914 (Rank 17/31). To address class imbalance, we used backtranslation for data augmentation, and cosine similarity to preserve label consistency after augmentation. This work emphasizes the need for hate speech detection in Devanagari-scripted languages and presents a foundation for further research.
♻ ☆ ProSparse: Introducing and Enhancing Intrinsic Activation Sparsity within Large Language Models
Activation sparsity refers to the existence of considerable weakly-contributed elements among activation outputs. As a prevalent property of the models using the ReLU activation function, activation sparsity has been proven a promising paradigm to boost model inference efficiency. Nevertheless, most large language models (LLMs) adopt activation functions without intrinsic activation sparsity (e.g., GELU and Swish). Some recent efforts have explored introducing ReLU or its variants as the substitutive activation function to help LLMs achieve activation sparsity and inference acceleration, but few can simultaneously obtain high sparsity and comparable model performance. This paper introduces a simple and effective sparsification method named "ProSparse" to push LLMs for higher activation sparsity while maintaining comparable performance. Specifically, after substituting the activation function of LLMs with ReLU, ProSparse adopts progressive sparsity regularization with a factor smoothly increasing along the multi-stage sine curves. This can enhance activation sparsity and mitigate performance degradation by avoiding radical shifts in activation distributions. With ProSparse, we obtain high sparsity of 89.32% for LLaMA2-7B, 88.80% for LLaMA2-13B, and 87.89% for end-size MiniCPM-1B, respectively, achieving comparable performance to their original Swish-activated versions. These present the most sparsely activated models among open-source LLaMA versions and competitive end-size models, considerably surpassing ReluLLaMA-7B (66.98%) and ReluLLaMA-13B (71.56%). Our inference acceleration experiments further demonstrate the significant practical acceleration potential of LLMs with higher activation sparsity, obtaining up to 4.52$\times$ inference speedup.
comment: 19 pages, 4 figures, 9 tables
♻ ☆ Missing Melodies: AI Music Generation and its "Nearly" Complete Omission of the Global South
Recent advances in generative AI have sparked renewed interest and expanded possibilities for music generation. However, the performance and versatility of these systems across musical genres are heavily influenced by the availability of training data. We conducted an extensive analysis of over one million hours of audio datasets used in AI music generation research and manually reviewed more than 200 papers from eleven prominent AI and music conferences and organizations (AAAI, ACM, EUSIPCO, EURASIP, ICASSP, ICML, IJCAI, ISMIR, NeurIPS, NIME, SMC) to identify a critical gap in the fair representation and inclusion of the musical genres of the Global South in AI research. Our findings reveal a stark imbalance: approximately 86% of the total dataset hours and over 93% of researchers focus primarily on music from the Global North. However, around 40% of these datasets include some form of non-Western music, genres from the Global South account for only 14.6% of the data. Furthermore, approximately 51% of the papers surveyed concentrate on symbolic music generation, a method that often fails to capture the cultural nuances inherent in music from regions such as South Asia, the Middle East, and Africa. As AI increasingly shapes the creation and dissemination of music, the significant underrepresentation of music genres in datasets and research presents a serious threat to global musical diversity. We also propose some important steps to mitigate these risks and foster a more inclusive future for AI-driven music generation.
comment: Submitted to CACM, 12 pages, 2 figures
♻ ☆ Large language models as oracles for instantiating ontologies with domain-specific knowledge
Background. Endowing intelligent systems with semantic data commonly requires designing and instantiating ontologies with domain-specific knowledge. Especially in the early phases, those activities are typically performed manually by human experts possibly leveraging on their own experience. The resulting process is therefore time-consuming, error-prone, and often biased by the personal background of the ontology designer. Objective. To mitigate that issue, we propose a novel domain-independent approach to automatically instantiate ontologies with domain-specific knowledge, by leveraging on large language models (LLMs) as oracles. Method. Starting from (i) an initial schema composed by inter-related classes and properties and (ii) a set of query templates, our method queries the LLM multiple times, and generates instances for both classes and properties from its replies. Thus, the ontology is automatically filled with domain-specific knowledge, compliant to the initial schema. As a result, the ontology is quickly and automatically enriched with manifold instances, which experts may consider to keep, adjust, discard, or complement according to their own needs and expertise. Contribution. We formalise our method in general way and instantiate it over various LLMs, as well as on a concrete case study. We report experiments rooted in the nutritional domain where an ontology of food meals and their ingredients is automatically instantiated from scratch, starting from a categorisation of meals and their relationships. There, we analyse the quality of the generated ontologies and compare ontologies attained by exploiting different LLMs. Experimentally, our approach achieves a quality metric that is up to five times higher than the state-of-the-art, while reducing erroneous entities and relations by up to ten times. Finally, we provide a SWOT analysis of the proposed method.
♻ ☆ UniBias: Unveiling and Mitigating LLM Bias through Internal Attention and FFN Manipulation NeurIPS 2024
Large language models (LLMs) have demonstrated impressive capabilities in various tasks using the in-context learning (ICL) paradigm. However, their effectiveness is often compromised by inherent bias, leading to prompt brittleness, i.e., sensitivity to design settings such as example selection, order, and prompt formatting. Previous studies have addressed LLM bias through external adjustment of model outputs, but the internal mechanisms that lead to such bias remain unexplored. Our work delves into these mechanisms, particularly investigating how feedforward neural networks (FFNs) and attention heads result in the bias of LLMs. By Interpreting the contribution of individual FFN vectors and attention heads, we identify the biased LLM components that skew LLMs' prediction toward specific labels. To mitigate these biases, we introduce UniBias, an inference-only method that effectively identifies and eliminates biased FFN vectors and attention heads. Extensive experiments across 12 NLP datasets demonstrate that UniBias significantly enhances ICL performance and alleviates prompt brittleness of LLMs.
comment: Accepted to NeurIPS 2024
♻ ☆ TCM-FTP: Fine-Tuning Large Language Models for Herbal Prescription Prediction
Traditional Chinese medicine (TCM) has relied on specific combinations of herbs in prescriptions to treat various symptoms and signs for thousands of years. Predicting TCM prescriptions poses a fascinating technical challenge with significant practical implications. However, this task faces limitations due to the scarcity of high-quality clinical datasets and the complex relationship between symptoms and herbs. To address these issues, we introduce \textit{DigestDS}, a novel dataset comprising practical medical records from experienced experts in digestive system diseases. We also propose a method, TCM-FTP (TCM Fine-Tuning Pre-trained), to leverage pre-trained large language models (LLMs) via supervised fine-tuning on \textit{DigestDS}. Additionally, we enhance computational efficiency using a low-rank adaptation technique. Moreover, TCM-FTP incorporates data augmentation by permuting herbs within prescriptions, exploiting their order-agnostic nature. Impressively, TCM-FTP achieves an F1-score of 0.8031, significantly outperforming previous methods. Furthermore, it demonstrates remarkable accuracy in dosage prediction, achieving a normalized mean square error of 0.0604. In contrast, LLMs without fine-tuning exhibit poor performance. Although LLMs have demonstrated wide-ranging capabilities, our work underscores the necessity of fine-tuning for TCM prescription prediction and presents an effective way to accomplish this.
comment: Camera-ready version to be published in BIBM 2024
♻ ☆ TouchTTS: An Embarrassingly Simple TTS Framework that Everyone Can Touch
It is well known that LLM-based systems are data-hungry. Recent LLM-based TTS works typically employ complex data processing pipelines to obtain high-quality training data. These sophisticated pipelines require excellent models at each stage (e.g., speech denoising, speech enhancement, speaker diarization, and punctuation models), which themselves demand high-quality training data and are rarely open-sourced. Even with state-of-the-art models, issues persist, such as incomplete background noise removal and misalignment between punctuation and actual speech pauses. Moreover, the stringent filtering strategies often retain only 10-30\% of the original data, significantly impeding data scaling efforts. In this work, we leverage a noise-robust audio tokenizer (S3Tokenizer) to design a simplified yet effective TTS data processing pipeline that maintains data quality while substantially reducing data acquisition costs, achieving a data retention rate of over 50\%. Beyond data scaling challenges, LLM-based TTS systems also incur higher deployment costs compared to conventional approaches. Current systems typically use LLMs solely for text-to-token generation, while requiring separate models (e.g., flow matching models) for token-to-waveform generation, which cannot be directly executed by LLM inference engines, further complicating deployment. To address these challenges, we eliminate redundant modules in both LLM and flow components, replacing the flow model backbone with an LLM architecture. Building upon this simplified flow backbone, we propose a unified architecture for both streaming and non-streaming inference, significantly reducing deployment costs. Finally, we explore the feasibility of unifying TTS and ASR tasks using the same data for training, thanks to the simplified pipeline and the S3Tokenizer that reduces the quality requirements for TTS training data.
comment: Technical Report
Information Retrieval 14
☆ Foundational Large Language Models for Materials Research
Materials discovery and development are critical for addressing global challenges. Yet, the exponential growth in materials science literature comprising vast amounts of textual data has created significant bottlenecks in knowledge extraction, synthesis, and scientific reasoning. Large Language Models (LLMs) offer unprecedented opportunities to accelerate materials research through automated analysis and prediction. Still, their effective deployment requires domain-specific adaptation for understanding and solving domain-relevant tasks. Here, we present LLaMat, a family of foundational models for materials science developed through continued pretraining of LLaMA models on an extensive corpus of materials literature and crystallographic data. Through systematic evaluation, we demonstrate that LLaMat excels in materials-specific NLP and structured information extraction while maintaining general linguistic capabilities. The specialized LLaMat-CIF variant demonstrates unprecedented capabilities in crystal structure generation, predicting stable crystals with high coverage across the periodic table. Intriguingly, despite LLaMA-3's superior performance in comparison to LLaMA-2, we observe that LLaMat-2 demonstrates unexpectedly enhanced domain-specific performance across diverse materials science tasks, including structured information extraction from text and tables, more particularly in crystal structure generation, a potential adaptation rigidity in overtrained LLMs. Altogether, the present work demonstrates the effectiveness of domain adaptation towards developing practically deployable LLM copilots for materials research. Beyond materials science, our findings reveal important considerations for domain adaptation of LLMs, such as model selection, training methodology, and domain-specific performance, which may influence the development of specialized scientific AI systems.
☆ SPRec: Leveraging Self-Play to Debias Preference Alignment for Large Language Model-based Recommendations
Large language models (LLMs) have attracted significant attention in recommendation systems. Current LLM-based recommender systems primarily rely on supervised fine-tuning (SFT) to train the model for recommendation tasks. However, relying solely on positive samples limits the model's ability to align with user satisfaction and expectations. To address this, researchers have introduced Direct Preference Optimization (DPO), which explicitly aligns recommendations with user preferences using offline preference ranking data. Despite its advantages, our theoretical analysis reveals that DPO inherently biases the model towards a few items, exacerbating the filter bubble issue and ultimately degrading user experience. In this paper, we propose SPRec, a novel self-play recommendation framework designed to mitigate over-recommendation and improve fairness without requiring additional data or manual intervention. In each self-play iteration, the model undergoes an SFT step followed by a DPO step, treating offline interaction data as positive samples and the predicted outputs from the previous iteration as negative samples. This effectively re-weights the DPO loss function using the model's logits, adaptively suppressing biased items. Extensive experiments on multiple real-world datasets demonstrate SPRec's effectiveness in enhancing recommendation accuracy and addressing fairness concerns.
☆ When Text Embedding Meets Large Language Model: A Comprehensive Survey
Text embedding has become a foundational technology in natural language processing (NLP) during the deep learning era, driving advancements across a wide array of downstream tasks. While many natural language understanding challenges can now be modeled using generative paradigms and leverage the robust generative and comprehension capabilities of large language models (LLMs), numerous practical applications, such as semantic matching, clustering, and information retrieval, continue to rely on text embeddings for their efficiency and effectiveness. In this survey, we categorize the interplay between LLMs and text embeddings into three overarching themes: (1) LLM-augmented text embedding, enhancing traditional embedding methods with LLMs; (2) LLMs as text embedders, utilizing their innate capabilities for embedding generation; and (3) Text embedding understanding with LLMs, leveraging LLMs to analyze and interpret embeddings. By organizing these efforts based on interaction patterns rather than specific downstream applications, we offer a novel and systematic overview of contributions from various research and application domains in the era of LLMs. Furthermore, we highlight the unresolved challenges that persisted in the pre-LLM era with pre-trained language models (PLMs) and explore the emerging obstacles brought forth by LLMs. Building on this analysis, we outline prospective directions for the evolution of text embedding, addressing both theoretical and practical opportunities in the rapidly advancing landscape of NLP.
comment: Work in progress
☆ Predicting Quality of Video Gaming Experience Using Global-Scale Telemetry Data and Federated Learning
Frames Per Second (FPS) significantly affects the gaming experience. Providing players with accurate FPS estimates prior to purchase benefits both players and game developers. However, we have a limited understanding of how to predict a game's FPS performance on a specific device. In this paper, we first conduct a comprehensive analysis of a wide range of factors that may affect game FPS on a global-scale dataset to identify the determinants of FPS. This includes player-side and game-side characteristics, as well as country-level socio-economic statistics. Furthermore, recognizing that accurate FPS predictions require extensive user data, which raises privacy concerns, we propose a federated learning-based model to ensure user privacy. Each player and game is assigned a unique learnable knowledge kernel that gradually extracts latent features for improved accuracy. We also introduce a novel training and prediction scheme that allows these kernels to be dynamically plug-and-play, effectively addressing cold start issues. To train this model with minimal bias, we collected a large telemetry dataset from 224 countries and regions, 100,000 users, and 835 games. Our model achieved a mean Wasserstein distance of 0.469 between predicted and ground truth FPS distributions, outperforming all baseline methods.
comment: 22 pages, 11 figures, 6 tables
☆ A Flexible Plug-and-Play Module for Generating Variable-Length
Deep supervised hashing has become a pivotal technique in large-scale image retrieval, offering significant benefits in terms of storage and search efficiency. However, existing deep supervised hashing models predominantly focus on generating fixed-length hash codes. This approach fails to address the inherent trade-off between efficiency and effectiveness when using hash codes of varying lengths. To determine the optimal hash code length for a specific task, multiple models must be trained for different lengths, leading to increased training time and computational overhead. Furthermore, the current paradigm overlooks the potential relationships between hash codes of different lengths, limiting the overall effectiveness of the models. To address these challenges, we propose the Nested Hash Layer (NHL), a plug-and-play module designed for existing deep supervised hashing models. The NHL framework introduces a novel mechanism to simultaneously generate hash codes of varying lengths in a nested manner. To tackle the optimization conflicts arising from the multiple learning objectives associated with different code lengths, we further propose an adaptive weights strategy that dynamically monitors and adjusts gradients during training. Additionally, recognizing that the structural information in longer hash codes can provide valuable guidance for shorter hash codes, we develop a long-short cascade self-distillation method within the NHL to enhance the overall quality of the generated hash codes. Extensive experiments demonstrate that NHL not only accelerates the training process but also achieves superior retrieval performance across various deep hashing models. Our code is publicly available at https://github.com/hly1998/NHL.
☆ Goal-Conditioned Supervised Learning for Multi-Objective Recommendation
Multi-objective learning endeavors to concurrently optimize multiple objectives using a single model, aiming to achieve high and balanced performance across these diverse objectives. However, it often involves a more complex optimization problem, particularly when navigating potential conflicts between objectives, leading to solutions with higher memory requirements and computational complexity. This paper introduces a Multi-Objective Goal-Conditioned Supervised Learning (MOGCSL) framework for automatically learning to achieve multiple objectives from offline sequential data. MOGCSL extends the conventional Goal-Conditioned Supervised Learning (GCSL) method to multi-objective scenarios by redefining goals from one-dimensional scalars to multi-dimensional vectors. The need for complex architectures and optimization constraints can be naturally eliminated. MOGCSL benefits from filtering out uninformative or noisy instances that do not achieve desirable long-term rewards. It also incorporates a novel goal-choosing algorithm to model and select "high" achievable goals for inference. While MOGCSL is quite general, we focus on its application to the next action prediction problem in commercial-grade recommender systems. In this context, any viable solution needs to be reasonably scalable and also be robust to large amounts of noisy data that is characteristic of this application space. We show that MOGCSL performs admirably on both counts. Specifically, extensive experiments conducted on real-world recommendation datasets validate its efficacy and efficiency. Also, analysis and experiments are included to explain its strength in discounting the noisier portions of training data in recommender systems.
☆ MOPI-HFRS: A Multi-objective Personalized Health-aware Food Recommendation System with LLM-enhanced Interpretation
The prevalence of unhealthy eating habits has become an increasingly concerning issue in the United States. However, major food recommendation platforms (e.g., Yelp) continue to prioritize users' dietary preferences over the healthiness of their choices. Although efforts have been made to develop health-aware food recommendation systems, the personalization of such systems based on users' specific health conditions remains under-explored. In addition, few research focus on the interpretability of these systems, which hinders users from assessing the reliability of recommendations and impedes the practical deployment of these systems. In response to this gap, we first establish two large-scale personalized health-aware food recommendation benchmarks at the first attempt. We then develop a novel framework, Multi-Objective Personalized Interpretable Health-aware Food Recommendation System (MOPI-HFRS), which provides food recommendations by jointly optimizing the three objectives: user preference, personalized healthiness and nutritional diversity, along with an large language model (LLM)-enhanced reasoning module to promote healthy dietary knowledge through the interpretation of recommended results. Specifically, this holistic graph learning framework first utilizes two structure learning and a structure pooling modules to leverage both descriptive features and health data. Then it employs Pareto optimization to achieve designed multi-facet objectives. Finally, to further promote the healthy dietary knowledge and awareness, we exploit an LLM by utilizing knowledge-infusion, prompting the LLMs with knowledge obtained from the recommendation model for interpretation.
♻ ☆ Logic Query of Thoughts: Guiding Large Language Models to Answer Complex Logic Queries with Knowledge Graphs
Despite the superb performance in many tasks, large language models (LLMs) bear the risk of generating hallucination or even wrong answers when confronted with tasks that demand the accuracy of knowledge. The issue becomes even more noticeable when addressing logic queries that require multiple logic reasoning steps. On the other hand, knowledge graph (KG) based question answering methods are capable of accurately identifying the correct answers with the help of knowledge graph, yet its accuracy could quickly deteriorate when the knowledge graph itself is sparse and incomplete. It remains a critical challenge on how to integrate knowledge graph reasoning with LLMs in a mutually beneficial way so as to mitigate both the hallucination problem of LLMs as well as the incompleteness issue of knowledge graphs. In this paper, we propose 'Logic-Query-of-Thoughts' (LGOT) which is the first of its kind to combine LLMs with knowledge graph based logic query reasoning. LGOT seamlessly combines knowledge graph reasoning and LLMs, effectively breaking down complex logic queries into easy to answer subquestions. Through the utilization of both knowledge graph reasoning and LLMs, it successfully derives answers for each subquestion. By aggregating these results and selecting the highest quality candidate answers for each step, LGOT achieves accurate results to complex questions. Our experimental findings demonstrate substantial performance enhancements, with up to 20% improvement over ChatGPT.
♻ ☆ HGCH: A Hyperbolic Graph Convolution Network Model for Heterogeneous Collaborative Graph Recommendation CIKM '24
User-item interaction data in collaborative filtering and graph modeling tasks often exhibit power-law characteristics, which suggest the suitability of hyperbolic space modeling. Hyperbolic Graph Convolution Neural Networks (HGCNs) are a novel technique that leverages the advantages of GCN and hyperbolic space, and then achieves remarkable results. However, existing HGCN methods have several drawbacks: they fail to fully leverage hyperbolic space properties due to arbitrary embedding initialization and imprecise tangent space aggregation; they overlook auxiliary information that could enrich the collaborative graph; and their training convergence is slow due to margin ranking loss and random negative sampling. To overcome these challenges, we propose Hyperbolic Graph Collaborative for Heterogeneous Recommendation (HGCH), an enhanced HGCN-based model for collaborative filtering that integrates diverse side information into a heterogeneous collaborative graph and improves training convergence speed. HGCH first preserves the long-tailed nature of the graph by initializing node embeddings with power law prior; then it aggregates neighbors in hyperbolic space using the gyromidpoint method for accurate computation; finally, it fuses multiple embeddings from different hyperbolic spaces by the gate fusion with prior. Moreover, HGCH employs a hyperbolic user-specific negative sampling to speed up convergence. We evaluate HGCH on four real datasets, and the results show that HGCH achieves competitive results and outperforms leading baselines, including HGCNs. Extensive ablation studies further confirm its effectiveness.
comment: Proceedings of the 33rd ACM International Conference on Information and Knowledge Management (CIKM '24)
♻ ☆ Large language models as oracles for instantiating ontologies with domain-specific knowledge
Background. Endowing intelligent systems with semantic data commonly requires designing and instantiating ontologies with domain-specific knowledge. Especially in the early phases, those activities are typically performed manually by human experts possibly leveraging on their own experience. The resulting process is therefore time-consuming, error-prone, and often biased by the personal background of the ontology designer. Objective. To mitigate that issue, we propose a novel domain-independent approach to automatically instantiate ontologies with domain-specific knowledge, by leveraging on large language models (LLMs) as oracles. Method. Starting from (i) an initial schema composed by inter-related classes and properties and (ii) a set of query templates, our method queries the LLM multiple times, and generates instances for both classes and properties from its replies. Thus, the ontology is automatically filled with domain-specific knowledge, compliant to the initial schema. As a result, the ontology is quickly and automatically enriched with manifold instances, which experts may consider to keep, adjust, discard, or complement according to their own needs and expertise. Contribution. We formalise our method in general way and instantiate it over various LLMs, as well as on a concrete case study. We report experiments rooted in the nutritional domain where an ontology of food meals and their ingredients is automatically instantiated from scratch, starting from a categorisation of meals and their relationships. There, we analyse the quality of the generated ontologies and compare ontologies attained by exploiting different LLMs. Experimentally, our approach achieves a quality metric that is up to five times higher than the state-of-the-art, while reducing erroneous entities and relations by up to ten times. Finally, we provide a SWOT analysis of the proposed method.
♻ ☆ Writing Style Matters: An Examination of Bias and Fairness in Information Retrieval Systems WSDM 25
The rapid advancement of Language Model technologies has opened new opportunities, but also introduced new challenges related to bias and fairness. This paper explores the uncharted territory of potential biases in state-of-the-art universal text embedding models towards specific document and query writing styles within Information Retrieval (IR) systems. Our investigation reveals that different embedding models exhibit different preferences of document writing style, while more informal and emotive styles are less favored by most embedding models. In terms of query writing styles, many embedding models tend to match the style of the query with the style of the retrieved documents, but some show a consistent preference for specific styles. Text embedding models fine-tuned on synthetic data generated by LLMs display a consistent preference for certain style of generated data. These biases in text embedding based IR systems can inadvertently silence or marginalize certain communication styles, thereby posing a significant threat to fairness in information retrieval. Finally, we also compare the answer styles of Retrieval Augmented Generation (RAG) systems based on different LLMs and find out that most text embedding models are biased towards LLM's answer styles when used as evaluation metrics for answer correctness. This study sheds light on the critical issue of writing style based bias in IR systems, offering valuable insights for the development of more fair and robust models.
comment: In Proceedings of the Eighteenth ACM International Conference on Web Search and Data Mining (WSDM 25)
♻ ☆ GANPrompt: Enhancing Robustness in LLM-Based Recommendations with GAN-Enhanced Diversity Prompts
In recent years, Large Language Models (LLMs) have demonstrated remarkable proficiency in comprehending and generating natural language, with a growing prevalence in the domain of recommendation systems. However, LLMs still face a significant challenge called prompt sensitivity, which refers to that it is highly susceptible to the influence of prompt words. This inconsistency in response to minor alterations in prompt input may compromise the accuracy and resilience of recommendation models. To address this issue, this paper proposes GANPrompt, a multi-dimensional LLMs prompt diversity framework based on Generative Adversarial Networks (GANs). The framework enhances the model's adaptability and stability to diverse prompts by integrating GANs generation techniques with the deep semantic understanding capabilities of LLMs. GANPrompt first trains a generator capable of producing diverse prompts by analysing multidimensional user behavioural data. These diverse prompts are then used to train the LLMs to improve its performance in the face of unseen prompts. Furthermore, to ensure a high degree of diversity and relevance of the prompts, this study introduces a mathematical theory-based diversity constraint mechanism that optimises the generated prompts to ensure that they are not only superficially distinct, but also semantically cover a wide range of user intentions. Through extensive experiments on multiple datasets, we demonstrate the effectiveness of the proposed framework, especially in improving the adaptability and robustness of recommendation systems in complex and dynamic environments. The experimental results demonstrate that GANPrompt yields substantial enhancements in accuracy and robustness relative to existing state-of-the-art methodologies.
♻ ☆ Task-level Distributionally Robust Optimization for Large Language Model-based Dense Retrieval AAAI25
Large Language Model-based Dense Retrieval (LLM-DR) optimizes over numerous heterogeneous fine-tuning collections from different domains. However, the discussion about its training data distribution is still minimal. Previous studies rely on empirically assigned dataset choices or sampling ratios, which inevitably lead to sub-optimal retrieval performances. In this paper, we propose a new task-level Distributionally Robust Optimization (tDRO) algorithm for LLM-DR fine-tuning, targeted at improving the universal domain generalization ability by end-to-end reweighting the data distribution of each task. The tDRO parameterizes the domain weights and updates them with scaled domain gradients. The optimized weights are then transferred to the LLM-DR fine-tuning to train more robust retrievers. Experiments show optimal improvements in large-scale retrieval benchmarks and reduce up to 30% dataset usage after applying our optimization algorithm with a series of different-sized LLM-DR models.
comment: Accepted by AAAI25. Source code is available at https://github.com/tdro-llm/tdro
♻ ☆ The Informational Role of Online Recommendations: Evidence from a Field Experiment
We conduct a field experiment on a movie-recommendation platform to investigate whether and how online recommendations influence consumption choices. Using a within-subjects design, our experiment measures the causal effect of recommendations on consumption and decomposes the relative importance of two economic mechanisms: expanding consumers' consideration sets and providing information about their idiosyncratic match value. We find that the informational component exerts a stronger influence - recommendations shape consumer beliefs, which in turn drive consumption, particularly among less experienced consumers. Our findings and experimental design provide valuable insights for the economic evaluation and optimisation of online recommendation systems.
Machine Learning 11
☆ Toward Foundation Model for Multivariate Wearable Sensing of Physiological Signals
Time-series foundation models have the ability to run inference, mainly forecasting, on any type of time series data, thanks to the informative representations comprising waveform features. Wearable sensing data, on the other hand, contain more variability in both patterns and frequency bands of interest and generally emphasize more on the ability to infer healthcare-related outcomes. The main challenge of crafting a foundation model for wearable sensing physiological signals is to learn generalizable representations that support efficient adaptation across heterogeneous sensing configurations and applications. In this work, we propose NormWear, a step toward such a foundation model, aiming to extract generalized and informative wearable sensing representations. NormWear has been pretrained on a large set of physiological signals, including PPG, ECG, EEG, GSR, and IMU, from various public resources. For a holistic assessment, we perform downstream evaluation on 11 public wearable sensing datasets, spanning 18 applications in the areas of mental health, body state inference, biomarker estimations, and disease risk evaluations. We demonstrate that NormWear achieves a better performance improvement over competitive baselines in general time series foundation modeling. In addition, leveraging a novel representation-alignment-match-based method, we align physiological signals embeddings with text embeddings. This alignment enables our proposed foundation model to perform zero-shot inference, allowing it to generalize to previously unseen wearable signal-based health applications. Finally, we perform nonlinear dynamic analysis on the waveform features extracted by the model at each intermediate layer. This analysis quantifies the model's internal processes, offering clear insights into its behavior and fostering greater trust in its inferences among end users.
comment: The code is available at: http://github.com/Mobile-Sensing-and-UbiComp-Laboratory/NormWear
☆ Towards joint graph and sampling set selection from data
We explore the problem of sampling graph signals in scenarios where the graph structure is not predefined and must be inferred from data. In this scenario, existing approaches rely on a two-step process, where a graph is learned first, followed by sampling. More generally, graph learning and graph signal sampling have been studied as two independent problems in the literature. This work provides a foundational step towards jointly optimizing the graph structure and sampling set. Our main contribution, Vertex Importance Sampling (VIS), is to show that the sampling set can be effectively determined from the vertex importance (node weights) obtained from graph learning. We further propose Vertex Importance Sampling with Repulsion (VISR), a greedy algorithm where spatially -separated "important" nodes are selected to ensure better reconstruction. Empirical results on simulated data show that sampling using VIS and VISR leads to competitive reconstruction performance and lower complexity than the conventional two-step approach of graph learning followed by graph sampling.
comment: 5 pages, 7 figures, IEEE Asilomar Conference on Signals, Systems, and Computers-2024
☆ A Quasilinear Algorithm for Computing Higher-Order Derivatives of Deep Feed-Forward Neural Networks
The use of neural networks for solving differential equations is practically difficult due to the exponentially increasing runtime of autodifferentiation when computing high-order derivatives. We propose $n$-TangentProp, the natural extension of the TangentProp formalism \cite{simard1991tangent} to arbitrarily many derivatives. $n$-TangentProp computes the exact derivative $d^n/dx^n f(x)$ in quasilinear, instead of exponential time, for a densely connected, feed-forward neural network $f$ with a smooth, parameter-free activation function. We validate our algorithm empirically across a range of depths, widths, and number of derivatives. We demonstrate that our method is particularly beneficial in the context of physics-informed neural networks where \ntp allows for significantly faster training times than previous methods and has favorable scaling with respect to both model size and loss-function complexity as measured by the number of required derivatives. The code for this paper can be found at https://github.com/kyrochi/n\_tangentprop.
comment: 11 pages, 10 figures
☆ New Approach to Clustering Random Attributes
This paper proposes a new method for similarity analysis and, consequently, a new algorithm for clustering different types of random attributes, both numerical and nominal. However, in order for nominal attributes to be clustered, their values must be properly encoded. In the encoding process, nominal attributes obtain a new representation in numerical form. Only the numeric attributes can be subjected to factor analysis, which allows them to be clustered in terms of their similarity to factors. The proposed method was tested for several sample datasets. It was found that the proposed method is universal. On the one hand, the method allows clustering of numerical attributes. On the other hand, it provides the ability to cluster nominal attributes. It also allows simultaneous clustering of numerical attributes and numerically encoded nominal attributes.
comment: 50 pages, 15 figures, 25 tables
☆ TelApart: Differentiating Network Faults from Customer-Premise Faults in Cable Broadband Networks
Two types of radio frequency (RF) impairments frequently occur in a cable broadband network: impairments that occur inside a cable network and impairments occur at the edge of the broadband network, i.e., in a subscriber's premise. Differentiating these two types of faults is important, as different faults require different types of technical personnel to repair them. Presently, the cable industry lacks publicly available tools to automatically diagnose the type of fault. In this work, we present TelApart, a fault diagnosis system for cable broadband networks. TelApart uses telemetry data collected by the Proactive Network Maintenance (PNM) infrastructure in cable networks to effectively differentiate the type of fault. Integral to TelApart's design is an unsupervised machine learning model that groups cable devices sharing similar anomalous patterns together. We use metrics derived from an ISP's customer trouble tickets to programmatically tune the model's hyper-parameters so that an ISP can deploy TelApart in various conditions without hand-tuning its hyper-parameters. We also address the data challenge that the telemetry data collected by the PNM system contain numerous missing, duplicated, and unaligned data points. Using real-world data contributed by a cable ISP, we show that TelApart can effectively identify different types of faults.
comment: 14 pages. arXiv admin note: text overlap with arXiv:2412.09564
♻ ☆ On the Workflows and Smells of Leaderboard Operations (LBOps): An Exploratory Study of Foundation Model Leaderboards
Foundation models (FM), such as large language models (LLMs), which are large-scale machine learning (ML) models, have demonstrated remarkable adaptability in various downstream software engineering (SE) tasks, such as code completion, code understanding, and software development. As a result, FM leaderboards have become essential tools for SE teams to compare and select the best third-party FMs for their specific products and purposes. However, the lack of standardized guidelines for FM evaluation and comparison threatens the transparency of FM leaderboards and limits stakeholders' ability to perform effective FM selection. As a first step towards addressing this challenge, our research focuses on understanding how these FM leaderboards operate in real-world scenarios ("leaderboard operations") and identifying potential pitfalls and areas for improvement ("leaderboard smells"). In this regard, we collect up to 1,045 FM leaderboards from five different sources: GitHub, Hugging Face Spaces, Papers With Code, spreadsheet and independent platform, to examine their documentation and engage in direct communication with leaderboard operators to understand their workflows. Through card sorting and negotiated agreement, we identify five distinct workflow patterns and develop a domain model that captures the key components and their interactions within these workflows. We then identify eight unique types of leaderboard smells in LBOps. By mitigating these smells, SE teams can improve transparency, accountability, and collaboration in current LBOps practices, fostering a more robust and responsible ecosystem for FM comparison and selection.
comment: awesome foundation model leaderboard list: https://github.com/SAILResearch/awesome-foundation-model-leaderboards
♻ ☆ Training LLMs over Neurally Compressed Text
In this paper, we explore the idea of training large language models (LLMs) over highly compressed text. While standard subword tokenizers compress text by a small factor, neural text compressors can achieve much higher rates of compression. If it were possible to train LLMs directly over neurally compressed text, this would confer advantages in training and serving efficiency, as well as easier handling of long text spans. The main obstacle to this goal is that strong compression tends to produce opaque outputs that are not well-suited for learning. In particular, we find that text na\"ively compressed via Arithmetic Coding is not readily learnable by LLMs. To overcome this, we propose Equal-Info Windows, a novel compression technique whereby text is segmented into blocks that each compress to the same bit length. Using this method, we demonstrate effective learning over neurally compressed text that improves with scale, and outperforms byte-level baselines by a wide margin on perplexity and inference speed benchmarks. While our method delivers worse perplexity than subword tokenizers for models trained with the same parameter count, it has the benefit of shorter sequence lengths. Shorter sequence lengths require fewer autoregressive generation steps, and reduce latency. Finally, we provide extensive analysis of the properties that contribute to learnability, and offer concrete suggestions for how to further improve the performance of high-compression tokenizers.
comment: Accepted in TMLR https://openreview.net/forum?id=pRvhMSV48t
♻ ☆ HYGENE: A Diffusion-based Hypergraph Generation Method
Hypergraphs are powerful mathematical structures that can model complex, high-order relationships in various domains, including social networks, bioinformatics, and recommender systems. However, generating realistic and diverse hypergraphs remains challenging due to their inherent complexity and lack of effective generative models. In this paper, we introduce a diffusion-based Hypergraph Generation (HYGENE) method that addresses these challenges through a progressive local expansion approach. HYGENE works on the bipartite representation of hypergraphs, starting with a single pair of connected nodes and iteratively expanding it to form the target hypergraph. At each step, nodes and hyperedges are added in a localized manner using a denoising diffusion process, which allows for the construction of the global structure before refining local details. Our experiments demonstrated the effectiveness of HYGENE, proving its ability to closely mimic a variety of properties in hypergraphs. To the best of our knowledge, this is the first attempt to employ deep learning models for hypergraph generation, and our work aims to lay the groundwork for future research in this area.
comment: arXiv admin note: text overlap with arXiv:2312.11529 by other authors
♻ ☆ Test Case-Informed Knowledge Tracing for Open-ended Coding Tasks
Open-ended coding tasks, which ask students to construct programs according to certain specifications, are common in computer science education. Student modeling can be challenging since their open-ended nature means that student code can be diverse. Traditional knowledge tracing (KT) models that only analyze response correctness may not fully capture nuances in student knowledge from student code. In this paper, we introduce Test case-Informed Knowledge Tracing for Open-ended Coding (TIKTOC), a framework to simultaneously analyze and predict both open-ended student code and whether the code passes each test case. We augment the existing CodeWorkout dataset with the test cases used for a subset of the open-ended coding questions, and propose a multi-task learning KT method to simultaneously analyze and predict 1) whether a student's code submission passes each test case and 2) the student's open-ended code, using a large language model as the backbone. We quantitatively show that these methods outperform existing KT methods for coding that only use the overall score a code submission receives. We also qualitatively demonstrate how test case information, combined with open-ended code, helps us gain fine-grained insights into student knowledge.
comment: Published in LAK 2025: The 15th International Learning Analytics and Knowledge Conference
♻ ☆ Interpretable Generalized Additive Models for Datasets with Missing Values NeurIPS 2024
Many important datasets contain samples that are missing one or more feature values. Maintaining the interpretability of machine learning models in the presence of such missing data is challenging. Singly or multiply imputing missing values complicates the model's mapping from features to labels. On the other hand, reasoning on indicator variables that represent missingness introduces a potentially large number of additional terms, sacrificing sparsity. We solve these problems with M-GAM, a sparse, generalized, additive modeling approach that incorporates missingness indicators and their interaction terms while maintaining sparsity through l0 regularization. We show that M-GAM provides similar or superior accuracy to prior methods while significantly improving sparsity relative to either imputation or naive inclusion of indicator variables.
comment: Published in NeurIPS 2024
♻ ☆ Conditioned quantum-assisted deep generative surrogate for particle-calorimeter interactions
Particle collisions at accelerators such as the Large Hadron Collider, recorded and analyzed by experiments such as ATLAS and CMS, enable exquisite measurements of the Standard Model and searches for new phenomena. Simulations of collision events at these detectors have played a pivotal role in shaping the design of future experiments and analyzing ongoing ones. However, the quest for accuracy in Large Hadron Collider (LHC) collisions comes at an imposing computational cost, with projections estimating the need for millions of CPU-years annually during the High Luminosity LHC (HL-LHC) run \cite{collaboration2022atlas}. Simulating a single LHC event with \textsc{Geant4} currently devours around 1000 CPU seconds, with simulations of the calorimeter subdetectors in particular imposing substantial computational demands \cite{rousseau2023experimental}. To address this challenge, we propose a conditioned quantum-assisted deep generative model. Our model integrates a conditioned variational autoencoder (VAE) on the exterior with a conditioned Restricted Boltzmann Machine (RBM) in the latent space, providing enhanced expressiveness compared to conventional VAEs. The RBM nodes and connections are meticulously engineered to enable the use of qubits and couplers on D-Wave's Pegasus-structured \textit{Advantage} quantum annealer (QA) for sampling. We introduce a novel method for conditioning the quantum-assisted RBM using \textit{flux biases}. We further propose a novel adaptive mapping to estimate the effective inverse temperature in quantum annealers. The effectiveness of our framework is illustrated using Dataset 2 of the CaloChallenge \cite{calochallenge}.
comment: 27 pages, 10 figures, 8 appendices
Artificial Intelligence 29
☆ Memory Layers at Scale
Memory layers use a trainable key-value lookup mechanism to add extra parameters to a model without increasing FLOPs. Conceptually, sparsely activated memory layers complement compute-heavy dense feed-forward layers, providing dedicated capacity to store and retrieve information cheaply. This work takes memory layers beyond proof-of-concept, proving their utility at contemporary scale. On downstream tasks, language models augmented with our improved memory layer outperform dense models with more than twice the computation budget, as well as mixture-of-expert models when matched for both compute and parameters. We find gains are especially pronounced for factual tasks. We provide a fully parallelizable memory layer implementation, demonstrating scaling laws with up to 128B memory parameters, pretrained to 1 trillion tokens, comparing to base models with up to 8B parameters.
☆ Congruence-based Learning of Probabilistic Deterministic Finite Automata
This work studies the question of learning probabilistic deterministic automata from language models. For this purpose, it focuses on analyzing the relations defined on algebraic structures over strings by equivalences and similarities on probability distributions. We introduce a congruence that extends the classical Myhill-Nerode congruence for formal languages. This new congruence is the basis for defining regularity over language models. We present an active learning algorithm that computes the quotient with respect to this congruence whenever the language model is regular. The paper also defines the notion of recognizability for language models and shows that it coincides with regularity for congruences. For relations which are not congruences, it shows that this is not the case. Finally, it discusses the impact of this result on learning in the context of language models.
☆ AI Red-Teaming is a Sociotechnical System. Now What?
As generative AI technologies find more and more real-world applications, the importance of testing their performance and safety seems paramount. ``Red-teaming'' has quickly become the primary approach to test AI models--prioritized by AI companies, and enshrined in AI policy and regulation. Members of red teams act as adversaries, probing AI systems to test their safety mechanisms and uncover vulnerabilities. Yet we know too little about this work and its implications. This essay calls for collaboration between computer scientists and social scientists to study the sociotechnical systems surrounding AI technologies, including the work of red-teaming, to avoid repeating the mistakes of the recent past. We highlight the importance of understanding the values and assumptions behind red-teaming, the labor involved, and the psychological impacts on red-teamers.
comment: 8 pages
☆ On Round-Off Errors and Gaussian Blur in Superresolution and in Image Registration
Superresolution theory and techniques seek to recover signals from samples in the presence of blur and noise. Discrete image registration can be an approach to fuse information from different sets of samples of the same signal. Quantization errors in the spatial domain are inherent to digital images. We consider superresolution and discrete image registration for one-dimensional spatially-limited piecewise constant functions which are subject to blur which is Gaussian or a mixture of Gaussians as well as to round-off errors. We describe a signal-dependent measurement matrix which captures both types of effects. For this setting we show that the difficulties in determining the discontinuity points from two sets of samples even in the absence of other types of noise. If the samples are also subject to statistical noise, then it is necessary to align and segment the data sequences to make the most effective inferences about the amplitudes and discontinuity points. Under some conditions on the blur, the noise, and the distance between discontinuity points, we prove that we can correctly align and determine the first samples following each discontinuity point in two data sequences with an approach based on dynamic programming.
☆ Let Curves Speak: A Continuous Glucose Monitor based Large Sensor Foundation Model for Diabetes Management
While previous studies of AI in diabetes management focus on long-term risk, research on near-future glucose prediction remains limited but important as it enables timely diabetes self-management. Integrating AI with continuous glucose monitoring (CGM) holds promise for near-future glucose prediction. However, existing models have limitations in capturing patterns of blood glucose fluctuations and demonstrate poor generalizability. A robust approach is needed to leverage massive CGM data for near-future glucose prediction. We propose large sensor models (LSMs) to capture knowledge in CGM data by modeling patients as sequences of glucose. CGM-LSM is pretrained on 15.96 million glucose records from 592 diabetes patients for near-future glucose prediction. We evaluated CGM-LSM against state-of-the-art methods using the OhioT1DM dataset across various metrics, prediction horizons, and unseen patients. Additionally, we assessed its generalizability across factors like diabetes type, age, gender, and hour of day. CGM-LSM achieved exceptional performance, with an rMSE of 29.81 mg/dL for type 1 diabetes patients and 23.49 mg/dL for type 2 diabetes patients in a two-hour prediction horizon. For the OhioT1DM dataset, CGM-LSM achieved a one-hour rMSE of 15.64 mg/dL, halving the previous best of 31.97 mg/dL. Robustness analyses revealed consistent performance not only for unseen patients and future periods, but also across diabetes type, age, and gender. The model demonstrated adaptability to different hours of day, maintaining accuracy across periods of various activity intensity levels. CGM-LSM represents a transformative step in diabetes management by leveraging pretraining to uncover latent glucose generation patterns in sensor data. Our findings also underscore the broader potential of LSMs to drive innovation across domains involving complex sensor data.
☆ The Unreasonable Effectiveness of Gaussian Score Approximation for Diffusion Models and its Applications
By learning the gradient of smoothed data distributions, diffusion models can iteratively generate samples from complex distributions. The learned score function enables their generalization capabilities, but how the learned score relates to the score of the underlying data manifold remains largely unclear. Here, we aim to elucidate this relationship by comparing learned neural scores to the scores of two kinds of analytically tractable distributions: Gaussians and Gaussian mixtures. The simplicity of the Gaussian model makes it theoretically attractive, and we show that it admits a closed-form solution and predicts many qualitative aspects of sample generation dynamics. We claim that the learned neural score is dominated by its linear (Gaussian) approximation for moderate to high noise scales, and supply both theoretical and empirical arguments to support this claim. Moreover, the Gaussian approximation empirically works for a larger range of noise scales than naive theory suggests it should, and is preferentially learned early in training. At smaller noise scales, we observe that learned scores are better described by a coarse-grained (Gaussian mixture) approximation of training data than by the score of the training distribution, a finding consistent with generalization. Our findings enable us to precisely predict the initial phase of trained models' sampling trajectories through their Gaussian approximations. We show that this allows the skipping of the first 15-30% of sampling steps while maintaining high sample quality (with a near state-of-the-art FID score of 1.93 on CIFAR-10 unconditional generation). This forms the foundation of a novel hybrid sampling method, termed analytical teleportation, which can seamlessly integrate with and accelerate existing samplers, including DPM-Solver-v3 and UniPC. Our findings suggest ways to improve the design and training of diffusion models.
comment: 69 pages, 34 figures. Published in TMLR. Previous shorter versions at arxiv.org/abs/2303.02490 and arxiv.org/abs/2311.10892
☆ TransferLight: Zero-Shot Traffic Signal Control on any Road-Network AAAI
Traffic signal control plays a crucial role in urban mobility. However, existing methods often struggle to generalize beyond their training environments to unseen scenarios with varying traffic dynamics. We present TransferLight, a novel framework designed for robust generalization across road-networks, diverse traffic conditions and intersection geometries. At its core, we propose a log-distance reward function, offering spatially-aware signal prioritization while remaining adaptable to varied lane configurations - overcoming the limitations of traditional pressure-based rewards. Our hierarchical, heterogeneous, and directed graph neural network architecture effectively captures granular traffic dynamics, enabling transferability to arbitrary intersection layouts. Using a decentralized multi-agent approach, global rewards, and novel state transition priors, we develop a single, weight-tied policy that scales zero-shot to any road network without re-training. Through domain randomization during training, we additionally enhance generalization capabilities. Experimental results validate TransferLight's superior performance in unseen scenarios, advancing practical, generalizable intelligent transportation systems to meet evolving urban traffic demands.
comment: AAAI Workshop Paper (MALTA)
☆ CUAL: Continual Uncertainty-aware Active Learner
AI deployed in many real-world use cases should be capable of adapting to novelties encountered after deployment. Here, we consider a challenging, under-explored and realistic continual adaptation problem: a deployed AI agent is continuously provided with unlabeled data that may contain not only unseen samples of known classes but also samples from novel (unknown) classes. In such a challenging setting, it has only a tiny labeling budget to query the most informative samples to help it continuously learn. We present a comprehensive solution to this complex problem with our model "CUAL" (Continual Uncertainty-aware Active Learner). CUAL leverages an uncertainty estimation algorithm to prioritize active labeling of ambiguous (uncertain) predicted novel class samples while also simultaneously pseudo-labeling the most certain predictions of each class. Evaluations across multiple datasets, ablations, settings and backbones (e.g. ViT foundation model) demonstrate our method's effectiveness. We will release our code upon acceptance.
☆ Doe-1: Closed-Loop Autonomous Driving with Large World Model
End-to-end autonomous driving has received increasing attention due to its potential to learn from large amounts of data. However, most existing methods are still open-loop and suffer from weak scalability, lack of high-order interactions, and inefficient decision-making. In this paper, we explore a closed-loop framework for autonomous driving and propose a large Driving wOrld modEl (Doe-1) for unified perception, prediction, and planning. We formulate autonomous driving as a next-token generation problem and use multi-modal tokens to accomplish different tasks. Specifically, we use free-form texts (i.e., scene descriptions) for perception and generate future predictions directly in the RGB space with image tokens. For planning, we employ a position-aware tokenizer to effectively encode action into discrete tokens. We train a multi-modal transformer to autoregressively generate perception, prediction, and planning tokens in an end-to-end and unified manner. Experiments on the widely used nuScenes dataset demonstrate the effectiveness of Doe-1 in various tasks including visual question-answering, action-conditioned video generation, and motion planning. Code: https://github.com/wzzheng/Doe.
comment: Code is available at: https://github.com/wzzheng/Doe
☆ TimeRefine: Temporal Grounding with Time Refining Video LLM
Video temporal grounding aims to localize relevant temporal boundaries in a video given a textual prompt. Recent work has focused on enabling Video LLMs to perform video temporal grounding via next-token prediction of temporal timestamps. However, accurately localizing timestamps in videos remains challenging for Video LLMs when relying solely on temporal token prediction. Our proposed TimeRefine addresses this challenge in two ways. First, instead of directly predicting the start and end timestamps, we reformulate the temporal grounding task as a temporal refining task: the model first makes rough predictions and then refines them by predicting offsets to the target segment. This refining process is repeated multiple times, through which the model progressively self-improves its temporal localization accuracy. Second, to enhance the model's temporal perception capabilities, we incorporate an auxiliary prediction head that penalizes the model more if a predicted segment deviates further from the ground truth, thus encouraging the model to make closer and more accurate predictions. Our plug-and-play method can be integrated into most LLM-based temporal grounding approaches. The experimental results demonstrate that TimeRefine achieves 3.6% and 5.0% mIoU improvements on the ActivityNet and Charades-STA datasets, respectively. Code and pretrained models will be released.
☆ Owl-1: Omni World Model for Consistent Long Video Generation
Video generation models (VGMs) have received extensive attention recently and serve as promising candidates for general-purpose large vision models. While they can only generate short videos each time, existing methods achieve long video generation by iteratively calling the VGMs, using the last-frame output as the condition for the next-round generation. However, the last frame only contains short-term fine-grained information about the scene, resulting in inconsistency in the long horizon. To address this, we propose an Omni World modeL (Owl-1) to produce long-term coherent and comprehensive conditions for consistent long video generation. As videos are observations of the underlying evolving world, we propose to model the long-term developments in a latent space and use VGMs to film them into videos. Specifically, we represent the world with a latent state variable which can be decoded into explicit video observations. These observations serve as a basis for anticipating temporal dynamics which in turn update the state variable. The interaction between evolving dynamics and persistent state enhances the diversity and consistency of the long videos. Extensive experiments show that Owl-1 achieves comparable performance with SOTA methods on VBench-I2V and VBench-Long, validating its ability to generate high-quality video observations. Code: https://github.com/huang-yh/Owl.
comment: Code is available at: https://github.com/huang-yh/Owl
☆ InternLM-XComposer2.5-OmniLive: A Comprehensive Multimodal System for Long-term Streaming Video and Audio Interactions
Creating AI systems that can interact with environments over long periods, similar to human cognition, has been a longstanding research goal. Recent advancements in multimodal large language models (MLLMs) have made significant strides in open-world understanding. However, the challenge of continuous and simultaneous streaming perception, memory, and reasoning remains largely unexplored. Current MLLMs are constrained by their sequence-to-sequence architecture, which limits their ability to process inputs and generate responses simultaneously, akin to being unable to think while perceiving. Furthermore, relying on long contexts to store historical data is impractical for long-term interactions, as retaining all information becomes costly and inefficient. Therefore, rather than relying on a single foundation model to perform all functions, this project draws inspiration from the concept of the Specialized Generalist AI and introduces disentangled streaming perception, reasoning, and memory mechanisms, enabling real-time interaction with streaming video and audio input. The proposed framework InternLM-XComposer2.5-OmniLive (IXC2.5-OL) consists of three key modules: (1) Streaming Perception Module: Processes multimodal information in real-time, storing key details in memory and triggering reasoning in response to user queries. (2) Multi-modal Long Memory Module: Integrates short-term and long-term memory, compressing short-term memories into long-term ones for efficient retrieval and improved accuracy. (3) Reasoning Module: Responds to queries and executes reasoning tasks, coordinating with the perception and memory modules. This project simulates human-like cognition, enabling multimodal large language models to provide continuous and adaptive service over time.
comment: Github Repo: https://github.com/InternLM/InternLM-XComposer/tree/main/InternLM-XComposer-2.5-OmniLive
☆ Neptune: The Long Orbit to Benchmarking Long Video Understanding
This paper describes a semi-automatic pipeline to generate challenging question-answer-decoy sets for understanding long videos. Many existing video datasets and models are focused on short clips (10s-30s). While some long video datasets do exist, they can often be solved by powerful image models applied per frame (and often to very few frames) in a video, and are usually manually annotated at high cost. In order to mitigate both these problems, we propose a scalable dataset creation pipeline which leverages large models (VLMs and LLMs), to automatically generate dense, time-aligned video captions, as well as tough question answer decoy sets for video segments (up to 15 minutes in length). Our dataset Neptune covers a broad range of long video reasoning abilities and consists of a subset that emphasizes multimodal reasoning. Since existing metrics for open-ended question answering are either rule-based or may rely on proprietary models, we provide a new open source model-based metric GEM to score open-ended responses on Neptune. Benchmark evaluations reveal that most current open-source long video models perform poorly on Neptune, particularly on questions testing temporal ordering, counting and state changes. Through Neptune, we aim to spur the development of more advanced models capable of understanding long videos. The dataset is available at https://github.com/google-deepmind/neptune
☆ A Theoretical Analysis of Soft-Label vs Hard-Label Training in Neural Networks
Knowledge distillation, where a small student model learns from a pre-trained large teacher model, has achieved substantial empirical success since the seminal work of \citep{hinton2015distilling}. Despite prior theoretical studies exploring the benefits of knowledge distillation, an important question remains unanswered: why does soft-label training from the teacher require significantly fewer neurons than directly training a small neural network with hard labels? To address this, we first present motivating experimental results using simple neural network models on a binary classification problem. These results demonstrate that soft-label training consistently outperforms hard-label training in accuracy, with the performance gap becoming more pronounced as the dataset becomes increasingly difficult to classify. We then substantiate these observations with a theoretical contribution based on two-layer neural network models. Specifically, we show that soft-label training using gradient descent requires only $O\left(\frac{1}{\gamma^2 \epsilon}\right)$ neurons to achieve a classification loss averaged over epochs smaller than some $\epsilon > 0$, where $\gamma$ is the separation margin of the limiting kernel. In contrast, hard-label training requires $O\left(\frac{1}{\gamma^4} \cdot \ln\left(\frac{1}{\epsilon}\right)\right)$ neurons, as derived from an adapted version of the gradient descent analysis in \citep{ji2020polylogarithmic}. This implies that when $\gamma \leq \epsilon$, i.e., when the dataset is challenging to classify, the neuron requirement for soft-label training can be significantly lower than that for hard-label training. Finally, we present experimental results on deep neural networks, further validating these theoretical findings.
comment: Main Body of the Paper is under Review at L4DC 2025
☆ DISHONEST: Dissecting misInformation Spread using Homogeneous sOcial NEtworks and Semantic Topic classification
The emergence of the COVID-19 pandemic resulted in a significant rise in the spread of misinformation on online platforms such as Twitter. Oftentimes this growth is blamed on the idea of the "echo chamber." However, the behavior said to characterize these echo chambers exists in two dimensions. The first is in a user's social interactions, where they are said to stick with the same clique of like-minded users. The second is in the content of their posts, where they are said to repeatedly espouse homogeneous ideas. In this study, we link the two by using Twitter's network of retweets to study social interactions and topic modeling to study tweet content. In order to measure the diversity of a user's interactions over time, we develop a novel metric to track the speed at which they travel through the social network. The application of these analysis methods to misinformation-focused data from the pandemic demonstrates correlation between social behavior and tweet content. We believe this correlation supports the common intuition about how antisocial users behave, and further suggests that it holds even in subcommunities already rife with misinformation.
☆ JuStRank: Benchmarking LLM Judges for System Ranking
Given the rapid progress of generative AI, there is a pressing need to systematically compare and choose between the numerous models and configurations available. The scale and versatility of such evaluations make the use of LLM-based judges a compelling solution for this challenge. Crucially, this approach requires first to validate the quality of the LLM judge itself. Previous work has focused on instance-based assessment of LLM judges, where a judge is evaluated over a set of responses, or response pairs, while being agnostic to their source systems. We argue that this setting overlooks critical factors affecting system-level ranking, such as a judge's positive or negative bias towards certain systems. To address this gap, we conduct the first large-scale study of LLM judges as system rankers. System scores are generated by aggregating judgment scores over multiple system outputs, and the judge's quality is assessed by comparing the resulting system ranking to a human-based ranking. Beyond overall judge assessment, our analysis provides a fine-grained characterization of judge behavior, including their decisiveness and bias.
☆ Sail into the Headwind: Alignment via Robust Rewards and Dynamic Labels against Reward Hacking
Aligning AI systems with human preferences typically suffers from the infamous reward hacking problem, where optimization of an imperfect reward model leads to undesired behaviors. In this paper, we investigate reward hacking in offline preference optimization, which aims to improve an initial model using a preference dataset. We identify two types of reward hacking stemming from statistical fluctuations in the dataset: Type I Reward Hacking due to subpar choices appearing more favorable, and Type II Reward Hacking due to decent choices appearing less favorable. We prove that many (mainstream or theoretical) preference optimization methods suffer from both types of reward hacking. To mitigate Type I Reward Hacking, we propose POWER, a new preference optimization method that combines Guiasu's weighted entropy with a robust reward maximization objective. POWER enjoys finite-sample guarantees under general function approximation, competing with the best covered policy in the data. To mitigate Type II Reward Hacking, we analyze the learning dynamics of preference optimization and develop a novel technique that dynamically updates preference labels toward certain "stationary labels", resulting in diminishing gradients for untrustworthy samples. Empirically, POWER with dynamic labels (POWER-DL) consistently outperforms state-of-the-art methods on alignment benchmarks, achieving improvements of up to 13.0 points on AlpacaEval 2.0 and 11.5 points on Arena-Hard over DPO, while also improving or maintaining performance on downstream tasks such as mathematical reasoning. Strong theoretical guarantees and empirical results demonstrate the promise of POWER-DL in mitigating reward hacking.
comment: 46 pages, 3 figures
☆ Systematic Analysis of LLM Contributions to Planning: Solver, Verifier, Heuristic
In this work, we provide a systematic analysis of how large language models (LLMs) contribute to solving planning problems. In particular, we examine how LLMs perform when they are used as problem solver, solution verifier, and heuristic guidance to improve intermediate solutions. Our analysis reveals that although it is difficult for LLMs to generate correct plans out-of-the-box, LLMs are much better at providing feedback signals to intermediate/incomplete solutions in the form of comparative heuristic functions. This evaluation framework provides insights into how future work may design better LLM-based tree-search algorithms to solve diverse planning and reasoning problems. We also propose a novel benchmark to evaluate LLM's ability to learn user preferences on the fly, which has wide applications in practical settings.
♻ ☆ HiRED: Attention-Guided Token Dropping for Efficient Inference of High-Resolution Vision-Language Models in Resource-Constrained Environments AAAI 2025
High-resolution Vision-Language Models (VLMs) have been widely used in multimodal tasks to enhance accuracy by preserving detailed image information. However, these models often generate excessive visual tokens due to encoding multiple partitions of the input image. Processing these excessive visual tokens is computationally challenging, especially in resource-constrained environments with commodity GPUs. To support high-resolution images while meeting resource constraints, we propose High-Resolution Early Dropping (HiRED), a token-dropping scheme that operates within a fixed token budget before the Large Language Model (LLM) stage. HiRED can be integrated with existing high-resolution VLMs in a plug-and-play manner, as it requires no additional training while still maintaining superior accuracy. We strategically use the vision encoder's attention in the initial layers to assess the visual content of each image partition and allocate the token budget accordingly. Then, using the attention in the final layer, we select the most important visual tokens from each partition within the allocated budget, dropping the rest. Empirically, when applied to LLaVA-Next-7B on NVIDIA TESLA P40 GPU, HiRED with a 20% token budget increases token generation throughput by 4.7, reduces first-token generation latency by 15 seconds, and saves 2.3 GB of GPU memory for a single inference. The code is available at https://github.com/hasanar1f/HiRED.
comment: Accepted in AAAI 2025
♻ ☆ Logic Query of Thoughts: Guiding Large Language Models to Answer Complex Logic Queries with Knowledge Graphs
Despite the superb performance in many tasks, large language models (LLMs) bear the risk of generating hallucination or even wrong answers when confronted with tasks that demand the accuracy of knowledge. The issue becomes even more noticeable when addressing logic queries that require multiple logic reasoning steps. On the other hand, knowledge graph (KG) based question answering methods are capable of accurately identifying the correct answers with the help of knowledge graph, yet its accuracy could quickly deteriorate when the knowledge graph itself is sparse and incomplete. It remains a critical challenge on how to integrate knowledge graph reasoning with LLMs in a mutually beneficial way so as to mitigate both the hallucination problem of LLMs as well as the incompleteness issue of knowledge graphs. In this paper, we propose 'Logic-Query-of-Thoughts' (LGOT) which is the first of its kind to combine LLMs with knowledge graph based logic query reasoning. LGOT seamlessly combines knowledge graph reasoning and LLMs, effectively breaking down complex logic queries into easy to answer subquestions. Through the utilization of both knowledge graph reasoning and LLMs, it successfully derives answers for each subquestion. By aggregating these results and selecting the highest quality candidate answers for each step, LGOT achieves accurate results to complex questions. Our experimental findings demonstrate substantial performance enhancements, with up to 20% improvement over ChatGPT.
♻ ☆ Conditioned quantum-assisted deep generative surrogate for particle-calorimeter interactions
Particle collisions at accelerators such as the Large Hadron Collider, recorded and analyzed by experiments such as ATLAS and CMS, enable exquisite measurements of the Standard Model and searches for new phenomena. Simulations of collision events at these detectors have played a pivotal role in shaping the design of future experiments and analyzing ongoing ones. However, the quest for accuracy in Large Hadron Collider (LHC) collisions comes at an imposing computational cost, with projections estimating the need for millions of CPU-years annually during the High Luminosity LHC (HL-LHC) run \cite{collaboration2022atlas}. Simulating a single LHC event with \textsc{Geant4} currently devours around 1000 CPU seconds, with simulations of the calorimeter subdetectors in particular imposing substantial computational demands \cite{rousseau2023experimental}. To address this challenge, we propose a conditioned quantum-assisted deep generative model. Our model integrates a conditioned variational autoencoder (VAE) on the exterior with a conditioned Restricted Boltzmann Machine (RBM) in the latent space, providing enhanced expressiveness compared to conventional VAEs. The RBM nodes and connections are meticulously engineered to enable the use of qubits and couplers on D-Wave's Pegasus-structured \textit{Advantage} quantum annealer (QA) for sampling. We introduce a novel method for conditioning the quantum-assisted RBM using \textit{flux biases}. We further propose a novel adaptive mapping to estimate the effective inverse temperature in quantum annealers. The effectiveness of our framework is illustrated using Dataset 2 of the CaloChallenge \cite{calochallenge}.
comment: 27 pages, 10 figures, 8 appendices
♻ ☆ LLMPhy: Complex Physical Reasoning Using Large Language Models and World Models
Physical reasoning is an important skill needed for robotic agents when operating in the real world. However, solving such reasoning problems often involves hypothesizing and reflecting over complex multi-body interactions under the effect of a multitude of physical forces and thus learning all such interactions poses a significant hurdle for state-of-the-art machine learning frameworks, including large language models (LLMs). To study this problem, we propose a new physical reasoning task and a dataset, dubbed TraySim. Our task involves predicting the dynamics of several objects on a tray that is given an external impact -- the domino effect of the ensued object interactions and their dynamics thus offering a challenging yet controlled setup, with the goal of reasoning being to infer the stability of the objects after the impact. To solve this complex physical reasoning task, we present LLMPhy, a zero-shot black-box optimization framework that leverages the physics knowledge and program synthesis abilities of LLMs, and synergizes these abilities with the world models built into modern physics engines. Specifically, LLMPhy uses an LLM to generate code to iteratively estimate the physical hyperparameters of the system (friction, damping, layout, etc.) via an implicit analysis-by-synthesis approach using a (non-differentiable) simulator in the loop and uses the inferred parameters to imagine the dynamics of the scene towards solving the reasoning task. To show the effectiveness of LLMPhy, we present experiments on our TraySim dataset to predict the steady-state poses of the objects. Our results show that the combination of the LLM and the physics engine leads to state-of-the-art zero-shot physical reasoning performance, while demonstrating superior convergence against standard black-box optimization methods and better estimation of the physical parameters.
♻ ☆ Limited but consistent gains in adversarial robustness by co-training object recognition models with human EEG ECCV
In contrast to human vision, artificial neural networks (ANNs) remain relatively susceptible to adversarial attacks. To address this vulnerability, efforts have been made to transfer inductive bias from human brains to ANNs, often by training the ANN representations to match their biological counterparts. Previous works relied on brain data acquired in rodents or primates using invasive techniques, from specific regions of the brain, under non-natural conditions (anesthetized animals), and with stimulus datasets lacking diversity and naturalness. In this work, we explored whether aligning model representations to human EEG responses to a rich set of real-world images increases robustness to ANNs. Specifically, we trained ResNet50-backbone models on a dual task of classification and EEG prediction; and evaluated their EEG prediction accuracy and robustness to adversarial attacks. We observed significant correlation between the networks' EEG prediction accuracy, often highest around 100 ms post stimulus onset, and their gains in adversarial robustness. Although effect size was limited, effects were consistent across different random initializations and robust for architectural variants. We further teased apart the data from individual EEG channels and observed strongest contribution from electrodes in the parieto-occipital regions. The demonstrated utility of human EEG for such tasks opens up avenues for future efforts that scale to larger datasets under diverse stimuli conditions with the promise of stronger effects.
comment: accepted as ECCV HCV workshop 2024 oral presentation
♻ ☆ Efficient Exploration and Discriminative World Model Learning with an Object-Centric Abstraction
In the face of difficult exploration problems in reinforcement learning, we study whether giving an agent an object-centric mapping (describing a set of items and their attributes) allow for more efficient learning. We found this problem is best solved hierarchically by modelling items at a higher level of state abstraction to pixels, and attribute change at a higher level of temporal abstraction to primitive actions. This abstraction simplifies the transition dynamic by making specific future states easier to predict. We make use of this to propose a fully model-based algorithm that learns a discriminative world model, plans to explore efficiently with only a count-based intrinsic reward, and can subsequently plan to reach any discovered (abstract) states. We demonstrate the model's ability to (i) efficiently solve single tasks, (ii) transfer zero-shot and few-shot across item types and environments, and (iii) plan across long horizons. Across a suite of 2D crafting and MiniHack environments, we empirically show our model significantly out-performs state-of-the-art low-level methods (without abstraction), as well as performant model-free and model-based methods using the same abstraction. Finally, we show how to learn low level object-perturbing policies via reinforcement learning, and the object mapping itself by supervised learning.
comment: Preprint. Additional results
♻ ☆ A Framework for testing Federated Learning algorithms using an edge-like environment
Federated Learning (FL) is a machine learning paradigm in which many clients cooperatively train a single centralized model while keeping their data private and decentralized. FL is commonly used in edge computing, which involves placing computer workloads (both hardware and software) as close as possible to the edge, where the data is being created and where actions are occurring, enabling faster response times, greater data privacy, and reduced data transfer costs. However, due to the heterogeneous data distributions/contents of clients, it is non-trivial to accurately evaluate the contributions of local models in global centralized model aggregation. This is an example of a major challenge in FL, commonly known as data imbalance or class imbalance. In general, testing and assessing FL algorithms can be a very difficult and complex task due to the distributed nature of the systems. In this work, a framework is proposed and implemented to assess FL algorithms in a more easy and scalable way. This framework is evaluated over a distributed edge-like environment managed by a container orchestration platform (i.e. Kubernetes).
♻ ☆ Generative Ghosts: Anticipating Benefits and Risks of AI Afterlives
As AI systems quickly improve in both breadth and depth of performance, they lend themselves to creating increasingly powerful and realistic agents, including the possibility of agents modeled on specific people. We anticipate that within our lifetimes it may become common practice for people to create custom AI agents to interact with loved ones and/or the broader world after death; indeed, the past year has seen a boom in startups purporting to offer such services. We call these generative ghosts, since such agents will be capable of generating novel content rather than merely parroting content produced by their creator while living. In this paper, we reflect on the history of technologies for AI afterlives, including current early attempts by individual enthusiasts and startup companies to create generative ghosts. We then introduce a novel design space detailing potential implementations of generative ghosts, and use this analytic framework to ground discussion of the practical and ethical implications of various approaches to designing generative ghosts, including potential positive and negative impacts on individuals and society. Based on these considerations, we lay out a research agenda for the AI and HCI research communities to better understand the risk/benefit landscape of this novel technology so as to ultimately empower people who wish to create and interact with AI afterlives to do so in a beneficial manner.
comment: version 4, updated to include new references and examples
♻ ☆ Towards One Model for Classical Dimensionality Reduction: A Probabilistic Perspective on UMAP and t-SNE
This paper shows that dimensionality reduction methods such as UMAP and t-SNE, can be approximately recast as MAP inference methods corresponding to a model introduced in ProbDR, that describes the graph Laplacian (an estimate for the precision/inverse covariance) matrix using a Wishart distribution, with a mean given by a non-linear covariance function evaluated on the latents. This interpretation offers deeper theoretical and semantic insights into such algorithms, by showing that variances corresponding to these covariances are low (and misspecified), and forging a connection to Gaussian process latent variable models by showing that well-known kernels can be used to describe covariances implied by graph Laplacians. We also introduce tools with which similar dimensionality reduction methods can be studied, and pose two areas of research arising from these interpretations.
comment: Updated preprint
♻ ☆ Piecing It All Together: Verifying Multi-Hop Multimodal Claims COLING 2025
Existing claim verification datasets often do not require systems to perform complex reasoning or effectively interpret multimodal evidence. To address this, we introduce a new task: multi-hop multimodal claim verification. This task challenges models to reason over multiple pieces of evidence from diverse sources, including text, images, and tables, and determine whether the combined multimodal evidence supports or refutes a given claim. To study this task, we construct MMCV, a large-scale dataset comprising 15k multi-hop claims paired with multimodal evidence, generated and refined using large language models, with additional input from human feedback. We show that MMCV is challenging even for the latest state-of-the-art multimodal large language models, especially as the number of reasoning hops increases. Additionally, we establish a human performance benchmark on a subset of MMCV. We hope this dataset and its evaluation task will encourage future research in multimodal multi-hop claim verification.
comment: COLING 2025
♻ ☆ Unveiling the Impact of Coding Data Instruction Fine-Tuning on Large Language Models Reasoning
Instruction Fine-Tuning (IFT) significantly enhances the zero-shot capabilities of pretrained Large Language Models (LLMs). While coding data is known to boost LLM reasoning abilities during pretraining, its role in activating internal reasoning capacities during IFT remains understudied. This paper investigates a key question: How does coding data impact LLMs' reasoning capacities during IFT stage? To explore this, we thoroughly examine the impact of coding data across different coding data proportions, model families, sizes, and reasoning domains, from various perspectives. Specifically, we create three IFT datasets with increasing coding data proportions, fine-tune six LLM backbones across different families and scales on these datasets, evaluate the tuned models' performance across twelve tasks in three reasoning domains, and analyze the outcomes from three broad-to-granular perspectives: overall, domain-level, and task-specific. Our holistic analysis provides valuable insights into each perspective. First, coding data tuning enhances the overall reasoning capabilities of LLMs across different model families and scales. Moreover, while the impact of coding data varies by domain, it shows consistent trends within each domain across different model families and scales. Additionally, coding data generally provides comparable task-specific benefits across model families, with optimal proportions in IFT datasets being task-dependent.
Computation and Language 93
☆ Fast Prompt Alignment for Text-to-Image Generation
Text-to-image generation has advanced rapidly, yet aligning complex textual prompts with generated visuals remains challenging, especially with intricate object relationships and fine-grained details. This paper introduces Fast Prompt Alignment (FPA), a prompt optimization framework that leverages a one-pass approach, enhancing text-to-image alignment efficiency without the iterative overhead typical of current methods like OPT2I. FPA uses large language models (LLMs) for single-iteration prompt paraphrasing, followed by fine-tuning or in-context learning with optimized prompts to enable real-time inference, reducing computational demands while preserving alignment fidelity. Extensive evaluations on the COCO Captions and PartiPrompts datasets demonstrate that FPA achieves competitive text-image alignment scores at a fraction of the processing time, as validated through both automated metrics (TIFA, VQA) and human evaluation. A human study with expert annotators further reveals a strong correlation between human alignment judgments and automated scores, underscoring the robustness of FPA's improvements. The proposed method showcases a scalable, efficient alternative to iterative prompt optimization, enabling broader applicability in real-time, high-demand settings. The codebase is provided to facilitate further research: https://github.com/tiktok/fast_prompt_alignment
comment: TikTok Technical Report
☆ Multimodal Latent Language Modeling with Next-Token Diffusion
Multimodal generative models require a unified approach to handle both discrete data (e.g., text and code) and continuous data (e.g., image, audio, video). In this work, we propose Latent Language Modeling (LatentLM), which seamlessly integrates continuous and discrete data using causal Transformers. Specifically, we employ a variational autoencoder (VAE) to represent continuous data as latent vectors and introduce next-token diffusion for autoregressive generation of these vectors. Additionally, we develop $\sigma$-VAE to address the challenges of variance collapse, which is crucial for autoregressive modeling. Extensive experiments demonstrate the effectiveness of LatentLM across various modalities. In image generation, LatentLM surpasses Diffusion Transformers in both performance and scalability. When integrated into multimodal large language models, LatentLM provides a general-purpose interface that unifies multimodal generation and understanding. Experimental results show that LatentLM achieves favorable performance compared to Transfusion and vector quantized models in the setting of scaling up training tokens. In text-to-speech synthesis, LatentLM outperforms the state-of-the-art VALL-E 2 model in speaker similarity and robustness, while requiring 10x fewer decoding steps. The results establish LatentLM as a highly effective and scalable approach to advance large multimodal models.
☆ Exploiting the Index Gradients for Optimization-Based Jailbreaking on Large Language Models
Despite the advancements in training Large Language Models (LLMs) with alignment techniques to enhance the safety of generated content, these models remain susceptible to jailbreak, an adversarial attack method that exposes security vulnerabilities in LLMs. Notably, the Greedy Coordinate Gradient (GCG) method has demonstrated the ability to automatically generate adversarial suffixes that jailbreak state-of-the-art LLMs. However, the optimization process involved in GCG is highly time-consuming, rendering the jailbreaking pipeline inefficient. In this paper, we investigate the process of GCG and identify an issue of Indirect Effect, the key bottleneck of the GCG optimization. To this end, we propose the Model Attack Gradient Index GCG (MAGIC), that addresses the Indirect Effect by exploiting the gradient information of the suffix tokens, thereby accelerating the procedure by having less computation and fewer iterations. Our experiments on AdvBench show that MAGIC achieves up to a 1.5x speedup, while maintaining Attack Success Rates (ASR) on par or even higher than other baselines. Our MAGIC achieved an ASR of 74% on the Llama-2 and an ASR of 54% when conducting transfer attacks on GPT-3.5. Code is available at https://github.com/jiah-li/magic.
comment: 13 pages,2 figures, accepted by The 31st International Conference on Computational Linguistics
☆ Der Effizienz- und Intelligenzbegriff in der Lexikographie und kuenstlichen Intelligenz: kann ChatGPT die lexikographische Textsorte nachbilden?
By means of pilot experiments for the language pair German and Galician, this paper examines the concept of efficiency and intelligence in lexicography and artificial intelligence, AI. The aim of the experiments is to gain empirically and statistically based insights into the lexicographical text type,dictionary article, in the responses of ChatGPT 3.5, as well as into the lexicographical data on which this chatbot was trained. Both quantitative and qualitative methods are used for this purpose. The analysis is based on the evaluation of the outputs of several sessions with the same prompt in ChatGPT 3.5. On the one hand, the algorithmic performance of intelligent systems is evaluated in comparison with data from lexicographical works. On the other hand, the ChatGPT data supplied is analysed using specific text passages of the aforementioned lexicographical text type. The results of this study not only help to evaluate the efficiency of this chatbot regarding the creation of dictionary articles, but also to delve deeper into the concept of intelligence, the thought processes and the actions to be carried out in both disciplines.
comment: 25 pages, in German language
☆ Advancing Single- and Multi-task Text Classification through Large Language Model Fine-tuning
Both encoder-only models (e.g., BERT, RoBERTa) and large language models (LLMs, e.g., Llama3) have been widely used for text classification tasks. However, there is a lack of systematic studies comparing the performance of encoder-based models and LLMs in text classification, particularly when fine-tuning is involved. This study employed a diverse range of models and methods, varying in size and architecture, and including both fine-tuned and pre-trained approaches. We first assessed the performances of these LLMs on the 20 Newsgroups (20NG) and MASSIVE datasets, comparing them to encoder-only RoBERTa models. Additionally, we explored the multi-task capabilities of both model types by combining multiple classification tasks, including intent detection and slot-filling, into a single model using data from both datasets. Our results indicate that fully fine-tuned Llama3-70B models outperform RoBERTa-large and other decoder LLMs across various classification tasks and datasets. Moreover, the consolidated multi-task fine-tuned LLMs matched the performance of dual-model setups in both tasks across both datasets. Overall, our study provides a comprehensive benchmark of encoder-only and LLM models on text classification tasks and demonstrates a method to combine two or more fully fine-tuned decoder LLMs for reduced latency and equivalent performance.
comment: 9 pages, 3 tables
☆ Machine Learning Information Retrieval and Summarisation to Support Systematic Review on Outcomes Based Contracting
As academic literature proliferates, traditional review methods are increasingly challenged by the sheer volume and diversity of available research. This article presents a study that aims to address these challenges by enhancing the efficiency and scope of systematic reviews in the social sciences through advanced machine learning (ML) and natural language processing (NLP) tools. In particular, we focus on automating stages within the systematic reviewing process that are time-intensive and repetitive for human annotators and which lend themselves to immediate scalability through tools such as information retrieval and summarisation guided by expert advice. The article concludes with a summary of lessons learnt regarding the integrated approach towards systematic reviews and future directions for improvement, including explainability.
☆ Can We Generate Visual Programs Without Prompting LLMs?
Visual programming prompts LLMs (large language mod-els) to generate executable code for visual tasks like visual question answering (VQA). Prompt-based methods are difficult to improve while also being unreliable and costly in both time and money. Our goal is to develop an efficient visual programming system without 1) using prompt-based LLMs at inference time and 2) a large set of program and answer annotations. We develop a synthetic data augmentation approach and alternative program generation method based on decoupling programs into higher-level skills called templates and the corresponding arguments. Our results show that with data augmentation, prompt-free smaller LLMs ($\approx$ 1B parameters) are competitive with state-of-the art models with the added benefit of much faster inference
☆ Bilevel Joint Unsupervised and Supervised Training for Automatic Speech Recognition
In this paper, we propose a bilevel joint unsupervised and supervised training (BL-JUST) framework for automatic speech recognition. Compared to the conventional pre-training and fine-tuning strategy which is a disconnected two-stage process, BL-JUST tries to optimize an acoustic model such that it simultaneously minimizes both the unsupervised and supervised loss functions. Because BL-JUST seeks matched local optima of both loss functions, acoustic representations learned by the acoustic model strike a good balance between being generic and task-specific. We solve the BL-JUST problem using penalty-based bilevel gradient descent and evaluate the trained deep neural network acoustic models on various datasets with a variety of architectures and loss functions. We show that BL-JUST can outperform the widely-used pre-training and fine-tuning strategy and some other popular semi-supervised techniques.
comment: Accepted by IEEE/ACM Transactions on Audio, Speech and Language Processing
☆ MaestroMotif: Skill Design from Artificial Intelligence Feedback
Describing skills in natural language has the potential to provide an accessible way to inject human knowledge about decision-making into an AI system. We present MaestroMotif, a method for AI-assisted skill design, which yields high-performing and adaptable agents. MaestroMotif leverages the capabilities of Large Language Models (LLMs) to effectively create and reuse skills. It first uses an LLM's feedback to automatically design rewards corresponding to each skill, starting from their natural language description. Then, it employs an LLM's code generation abilities, together with reinforcement learning, for training the skills and combining them to implement complex behaviors specified in language. We evaluate MaestroMotif using a suite of complex tasks in the NetHack Learning Environment (NLE), demonstrating that it surpasses existing approaches in both performance and usability.
☆ TECO: Improving Multimodal Intent Recognition with Text Enhancement through Commonsense Knowledge Extraction ACL
The objective of multimodal intent recognition (MIR) is to leverage various modalities-such as text, video, and audio-to detect user intentions, which is crucial for understanding human language and context in dialogue systems. Despite advances in this field, two main challenges persist: (1) effectively extracting and utilizing semantic information from robust textual features; (2) aligning and fusing non-verbal modalities with verbal ones effectively. This paper proposes a Text Enhancement with CommOnsense Knowledge Extractor (TECO) to address these challenges. We begin by extracting relations from both generated and retrieved knowledge to enrich the contextual information in the text modality. Subsequently, we align and integrate visual and acoustic representations with these enhanced text features to form a cohesive multimodal representation. Our experimental results show substantial improvements over existing baseline methods.
comment: Accepted at PACLIC 2024
☆ Continual Learning for Encoder-only Language Models via a Discrete Key-Value Bottleneck
Continual learning remains challenging across various natural language understanding tasks. When models are updated with new training data, they risk catastrophic forgetting of prior knowledge. In the present work, we introduce a discrete key-value bottleneck for encoder-only language models, allowing for efficient continual learning by requiring only localized updates. Inspired by the success of a discrete key-value bottleneck in vision, we address new and NLP-specific challenges. We experiment with different bottleneck architectures to find the most suitable variants regarding language, and present a generic discrete key initialization technique for NLP that is task independent. We evaluate the discrete key-value bottleneck in four continual learning NLP scenarios and demonstrate that it alleviates catastrophic forgetting. We showcase that it offers competitive performance to other popular continual learning methods, with lower computational costs.
☆ EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance
As large language models (LLMs) continue to advance, the demand for higher quality and faster processing of long contexts across various applications is growing. KV cache is widely adopted as it stores previously generated key and value tokens, effectively reducing redundant computations during inference. However, as memory overhead becomes a significant concern, efficient compression of KV cache has gained increasing attention. Most existing methods perform compression from two perspectives: identifying important tokens and designing compression strategies. However, these approaches often produce biased distributions of important tokens due to the influence of accumulated attention scores or positional encoding. Furthermore, they overlook the sparsity and redundancy across different heads, which leads to difficulties in preserving the most effective information at the head level. To this end, we propose EMS to overcome these limitations, while achieving better KV cache compression under extreme compression ratios. Specifically, we introduce a Global-Local score that combines accumulated attention scores from both global and local KV tokens to better identify the token importance. For the compression strategy, we design an adaptive and unified Evict-then-Merge framework that accounts for the sparsity and redundancy of KV tokens across different heads. Additionally, we implement the head-wise parallel compression through a zero-class mechanism to enhance efficiency. Extensive experiments demonstrate our SOTA performance even under extreme compression ratios. EMS consistently achieves the lowest perplexity, improves scores by over 1.28 points across four LLMs on LongBench under a 256 cache budget, and preserves 95% retrieval accuracy with a cache budget less than 2% of the context length in the Needle-in-a-Haystack task.
☆ GR-NLP-TOOLKIT: An Open-Source NLP Toolkit for Modern Greek COLING 2025
We present GR-NLP-TOOLKIT, an open-source natural language processing (NLP) toolkit developed specifically for modern Greek. The toolkit provides state-of-the-art performance in five core NLP tasks, namely part-of-speech tagging, morphological tagging, dependency parsing, named entity recognition, and Greeklishto-Greek transliteration. The toolkit is based on pre-trained Transformers, it is freely available, and can be easily installed in Python (pip install gr-nlp-toolkit). It is also accessible through a demonstration platform on HuggingFace, along with a publicly available API for non-commercial use. We discuss the functionality provided for each task, the underlying methods, experiments against comparable open-source toolkits, and future possible enhancements. The toolkit is available at: https://github.com/nlpaueb/gr-nlp-toolkit
comment: Accepted Demo Paper @ COLING 2025 (Github: https://github.com/nlpaueb/gr-nlp-toolkit/, Demo: https://huggingface.co/spaces/AUEB-NLP/greek-nlp-toolkit-demo, API: https://huggingface.co/spaces/AUEB-NLP/The-Greek-NLP-API)
☆ Bridging Relevance and Reasoning: Rationale Distillation in Retrieval-Augmented Generation
The reranker and generator are two critical components in the Retrieval-Augmented Generation (i.e., RAG) pipeline, responsible for ranking relevant documents and generating responses. However, due to differences in pre-training data and objectives, there is an inevitable gap between the documents ranked as relevant by the reranker and those required by the generator to support answering the query. To address this gap, we propose RADIO, a novel and practical preference alignment framework with RAtionale DIstillatiOn. Specifically, We first propose a rationale extraction method that leverages the reasoning capabilities of Large Language Models (LLMs) to extract the rationales necessary for answering the query. Subsequently, a rationale-based alignment process is designed to rerank the documents based on the extracted rationales, and fine-tune the reranker to align the preferences. We conduct extensive experiments on two tasks across three datasets to demonstrate the effectiveness of our approach compared to baseline methods. Our code is released online to ease reproduction.
comment: under review
☆ Comparative Opinion Mining in Product Reviews: Multi-perspective Prompt-based Learning
Comparative reviews are pivotal in understanding consumer preferences and influencing purchasing decisions. Comparative Quintuple Extraction (COQE) aims to identify five key components in text: the target entity, compared entities, compared aspects, opinions on these aspects, and polarity. Extracting precise comparative information from product reviews is challenging due to nuanced language and sequential task errors in traditional methods. To mitigate these problems, we propose MTP-COQE, an end-to-end model designed for COQE. Leveraging multi-perspective prompt-based learning, MTP-COQE effectively guides the generative model in comparative opinion mining tasks. Evaluation on the Camera-COQE (English) and VCOM (Vietnamese) datasets demonstrates MTP-COQE's efficacy in automating COQE, achieving superior performance with a 1.41% higher F1 score than the previous baseline models on the English dataset. Additionally, we designed a strategy to limit the generative model's creativity to ensure the output meets expectations. We also performed data augmentation to address data imbalance and to prevent the model from becoming biased towards dominant samples.
☆ Multi-perspective Alignment for Increasing Naturalness in Neural Machine Translation
Neural machine translation (NMT) systems amplify lexical biases present in their training data, leading to artificially impoverished language in output translations. These language-level characteristics render automatic translations different from text originally written in a language and human translations, which hinders their usefulness in for example creating evaluation datasets. Attempts to increase naturalness in NMT can fall short in terms of content preservation, where increased lexical diversity comes at the cost of translation accuracy. Inspired by the reinforcement learning from human feedback framework, we introduce a novel method that rewards both naturalness and content preservation. We experiment with multiple perspectives to produce more natural translations, aiming at reducing machine and human translationese. We evaluate our method on English-to-Dutch literary translation, and find that our best model produces translations that are lexically richer and exhibit more properties of human-written language, without loss in translation accuracy.
☆ Bootstrapping Language-Guided Navigation Learning with Self-Refining Data Flywheel
Creating high-quality data for training robust language-instructed agents is a long-lasting challenge in embodied AI. In this paper, we introduce a Self-Refining Data Flywheel (SRDF) that generates high-quality and large-scale navigational instruction-trajectory pairs by iteratively refining the data pool through the collaboration between two models, the instruction generator and the navigator, without any human-in-the-loop annotation. Specifically, SRDF starts with using a base generator to create an initial data pool for training a base navigator, followed by applying the trained navigator to filter the data pool. This leads to higher-fidelity data to train a better generator, which can, in turn, produce higher-quality data for training the next-round navigator. Such a flywheel establishes a data self-refining process, yielding a continuously improved and highly effective dataset for large-scale language-guided navigation learning. Our experiments demonstrate that after several flywheel rounds, the navigator elevates the performance boundary from 70% to 78% SPL on the classic R2R test set, surpassing human performance (76%) for the first time. Meanwhile, this process results in a superior generator, evidenced by a SPICE increase from 23.5 to 26.2, better than all previous VLN instruction generation methods. Finally, we demonstrate the scalability of our method through increasing environment and instruction diversity, and the generalization ability of our pre-trained navigator across various downstream navigation tasks, surpassing state-of-the-art methods by a large margin in all cases.
comment: 28 pages, Code and data are available at https://github.com/wz0919/VLN-SRDF
☆ Mitigating Out-of-Entity Errors in Named Entity Recognition: A Sentence-Level Strategy COLING 2025
Many previous models of named entity recognition (NER) suffer from the problem of Out-of-Entity (OOE), i.e., the tokens in the entity mentions of the test samples have not appeared in the training samples, which hinders the achievement of satisfactory performance. To improve OOE-NER performance, in this paper, we propose a new framework, namely S+NER, which fully leverages sentence-level information. Our S+NER achieves better OOE-NER performance mainly due to the following two particular designs. 1) It first exploits the pre-trained language model's capability of understanding the target entity's sentence-level context with a template set. 2) Then, it refines the sentence-level representation based on the positive and negative templates, through a contrastive learning strategy and template pooling method, to obtain better NER results. Our extensive experiments on five benchmark datasets have demonstrated that, our S+NER outperforms some state-of-the-art OOE-NER models.
comment: Accepted by COLING 2025
☆ Assessing Personalized AI Mentoring with Large Language Models in the Computing Field
This paper provides an in-depth evaluation of three state-of-the-art Large Language Models (LLMs) for personalized career mentoring in the computing field, using three distinct student profiles that consider gender, race, and professional levels. We evaluated the performance of GPT-4, LLaMA 3, and Palm 2 using a zero-shot learning approach without human intervention. A quantitative evaluation was conducted through a custom natural language processing analytics pipeline to highlight the uniqueness of the responses and to identify words reflecting each student's profile, including race, gender, or professional level. The analysis of frequently used words in the responses indicates that GPT-4 offers more personalized mentoring compared to the other two LLMs. Additionally, a qualitative evaluation was performed to see if human experts reached similar conclusions. The analysis of survey responses shows that GPT-4 outperformed the other two LLMs in delivering more accurate and useful mentoring while addressing specific challenges with encouragement languages. Our work establishes a foundation for developing personalized mentoring tools based on LLMs, incorporating human mentors in the process to deliver a more impactful and tailored mentoring experience.
Detecting Conversational Mental Manipulation with Intent-Aware Prompting
Mental manipulation severely undermines mental wellness by covertly and negatively distorting decision-making. While there is an increasing interest in mental health care within the natural language processing community, progress in tackling manipulation remains limited due to the complexity of detecting subtle, covert tactics in conversations. In this paper, we propose Intent-Aware Prompting (IAP), a novel approach for detecting mental manipulations using large language models (LLMs), providing a deeper understanding of manipulative tactics by capturing the underlying intents of participants. Experimental results on the MentalManip dataset demonstrate superior effectiveness of IAP against other advanced prompting strategies. Notably, our approach substantially reduces false negatives, helping detect more instances of mental manipulation with minimal misjudgment of positive cases. The code of this paper is available at https://github.com/Anton-Jiayuan-MA/Manip-IAP.
☆ Learning to Reason via Self-Iterative Process Feedback for Small Language Models COLING 2025
Small language models (SLMs) are more efficient, cost-effective, and customizable than large language models (LLMs), though they often underperform in specific areas like reasoning. Past methods for enhancing SLMs' reasoning, such as supervised fine-tuning and distillation, often depend on costly external signals, resulting in SLMs being overly confident with limited supervision signals, thus limiting their abilities. Therefore, this study enables SLMs to learn to reason from self-iterative feedback. By combining odds ratio preference optimization (ORPO), we fine-tune and align SLMs using positive and negative signals generated by themselves. Additionally, we introduce process supervision for rewards in preference alignment by sampling-based inference simulation and process reward models. Compared to Supervised Fine-Tuning (SFT), our method improves the performance of Gemma-2B by 12.43 (Acc) on GSM8K and 3.95 (Pass@1) on MBPP. Furthermore, the proposed method also demonstrated superior out-of-domain generalization capabilities on MMLU_Math and HumanEval.
comment: Accepted by COLING 2025
☆ The Roles of English in Evaluating Multilingual Language Models
Multilingual natural language processing is getting increased attention, with numerous models, benchmarks, and methods being released for many languages. English is often used in multilingual evaluation to prompt language models (LMs), mainly to overcome the lack of instruction tuning data in other languages. In this position paper, we lay out two roles of English in multilingual LM evaluations: as an interface and as a natural language. We argue that these roles have different goals: task performance versus language understanding. This discrepancy is highlighted with examples from datasets and evaluation setups. Numerous works explicitly use English as an interface to boost task performance. We recommend to move away from this imprecise method and instead focus on furthering language understanding.
comment: NoDaLiDa 2025
☆ SweetieChat: A Strategy-Enhanced Role-playing Framework for Diverse Scenarios Handling Emotional Support Agent COLING 2025
Large Language Models (LLMs) have demonstrated promising potential in providing empathetic support during interactions. However, their responses often become verbose or overly formulaic, failing to adequately address the diverse emotional support needs of real-world scenarios. To tackle this challenge, we propose an innovative strategy-enhanced role-playing framework, designed to simulate authentic emotional support conversations. Specifically, our approach unfolds in two steps: (1) Strategy-Enhanced Role-Playing Interactions, which involve three pivotal roles -- Seeker, Strategy Counselor, and Supporter -- engaging in diverse scenarios to emulate real-world interactions and promote a broader range of dialogues; and (2) Emotional Support Agent Training, achieved through fine-tuning LLMs using our specially constructed dataset. Within this framework, we develop the \textbf{ServeForEmo} dataset, comprising an extensive collection of 3.7K+ multi-turn dialogues and 62.8K+ utterances. We further present \textbf{SweetieChat}, an emotional support agent capable of handling diverse open-domain scenarios. Extensive experiments and human evaluations confirm the framework's effectiveness in enhancing emotional support, highlighting its unique ability to provide more nuanced and tailored assistance.
comment: 24 pages. Accepted by COLING 2025
☆ NyayaAnumana & INLegalLlama: The Largest Indian Legal Judgment Prediction Dataset and Specialized Language Model for Enhanced Decision Analysis COLING 2025
The integration of artificial intelligence (AI) in legal judgment prediction (LJP) has the potential to transform the legal landscape, particularly in jurisdictions like India, where a significant backlog of cases burdens the legal system. This paper introduces NyayaAnumana, the largest and most diverse corpus of Indian legal cases compiled for LJP, encompassing a total of 7,02,945 preprocessed cases. NyayaAnumana, which combines the words "Nyay" (judgment) and "Anuman" (prediction or inference) respectively for most major Indian languages, includes a wide range of cases from the Supreme Court, High Courts, Tribunal Courts, District Courts, and Daily Orders and, thus, provides unparalleled diversity and coverage. Our dataset surpasses existing datasets like PredEx and ILDC, offering a comprehensive foundation for advanced AI research in the legal domain. In addition to the dataset, we present INLegalLlama, a domain-specific generative large language model (LLM) tailored to the intricacies of the Indian legal system. It is developed through a two-phase training approach over a base LLaMa model. First, Indian legal documents are injected using continual pretraining. Second, task-specific supervised finetuning is done. This method allows the model to achieve a deeper understanding of legal contexts. Our experiments demonstrate that incorporating diverse court data significantly boosts model accuracy, achieving approximately 90% F1-score in prediction tasks. INLegalLlama not only improves prediction accuracy but also offers comprehensible explanations, addressing the need for explainability in AI-assisted legal decisions.
comment: Accepted on COLING 2025
☆ SmolTulu: Higher Learning Rate to Batch Size Ratios Can Lead to Better Reasoning in SLMs
We present SmolTulu-1.7b-Instruct, referenced in this report as SmolTulu-DPO-1130, an instruction-tuned language model that adapts AllenAI's Tulu 3 post-training pipeline to enhance Huggingface's SmolLM2-1.7B base model. Through comprehensive empirical analysis using a 135M parameter model, we demonstrate that the relationship between learning rate and batch size significantly impacts model performance in a task-dependent manner. Our findings reveal a clear split: reasoning tasks like ARC and GSM8K benefit from higher learning rate to batch size ratios, while pattern recognition tasks such as HellaSwag and IFEval show optimal performance with lower ratios. These insights informed the development of SmolTulu, which achieves state-of-the-art performance among sub-2B parameter models on instruction following, scoring 67.7% on IFEval ($\Delta$11%), and mathematical reasoning with 51.6% on GSM8K ($\Delta$3.4%), with an alternate version achieving scoring 57.1% on ARC ($\Delta5.4%$). We release our model, training recipes, and ablation studies to facilitate further research in efficient model alignment, demonstrating that careful adaptation of optimization dynamics can help bridge the capability gap between small and large language models.
comment: 10 pages, 4 figures, and 13 tables. For the SmolTulu-1.7b-instruct model, see: https://huggingface.co/SultanR/SmolTulu-1.7b-Instruct
☆ BEIR-NL: Zero-shot Information Retrieval Benchmark for the Dutch Language COLING
Zero-shot evaluation of information retrieval (IR) models is often performed using BEIR; a large and heterogeneous benchmark composed of multiple datasets, covering different retrieval tasks across various domains. Although BEIR has become a standard benchmark for the zero-shot setup, its exclusively English content reduces its utility for underrepresented languages in IR, including Dutch. To address this limitation and encourage the development of Dutch IR models, we introduce BEIR-NL by automatically translating the publicly accessible BEIR datasets into Dutch. Using BEIR-NL, we evaluated a wide range of multilingual dense ranking and reranking models, as well as the lexical BM25 method. Our experiments show that BM25 remains a competitive baseline, and is only outperformed by the larger dense models trained for retrieval. When combined with reranking models, BM25 achieves performance on par with the best dense ranking models. In addition, we explored the impact of translation on the data by back-translating a selection of datasets to English, and observed a performance drop for both dense and lexical methods, indicating the limitations of translation for creating benchmarks. BEIR-NL is publicly available on the Hugging Face hub.
comment: To be presented at BUCC 2025 (COLING)
☆ Large Language Models Still Face Challenges in Multi-Hop Reasoning with External Knowledge
We carry out a series of experiments to test large language models' multi-hop reasoning ability from three aspects: selecting and combining external knowledge, dealing with non-sequential reasoning tasks and generalising to data samples with larger numbers of hops. We test the GPT-3.5 model on four reasoning benchmarks with Chain-of-Thought prompting (and its variations). Our results reveal that despite the amazing performance achieved by large language models on various reasoning tasks, models still suffer from severe drawbacks which shows a large gap with humans.
☆ Rumor Detection on Social Media with Temporal Propagation Structure Optimization COLING'25
Traditional methods for detecting rumors on social media primarily focus on analyzing textual content, often struggling to capture the complexity of online interactions. Recent research has shifted towards leveraging graph neural networks to model the hierarchical conversation structure that emerges during rumor propagation. However, these methods tend to overlook the temporal aspect of rumor propagation and may disregard potential noise within the propagation structure. In this paper, we propose a novel approach that incorporates temporal information by constructing a weighted propagation tree, where the weight of each edge represents the time interval between connected posts. Drawing upon the theory of structural entropy, we transform this tree into a coding tree. This transformation aims to preserve the essential structure of rumor propagation while reducing noise. Finally, we introduce a recursive neural network to learn from the coding tree for rumor veracity prediction. Experimental results on two common datasets demonstrate the superiority of our approach.
comment: COLING'25
Code LLMs: A Taxonomy-based Survey
Large language models (LLMs) have demonstrated remarkable capabilities across various NLP tasks and have recently expanded their impact to coding tasks, bridging the gap between natural languages (NL) and programming languages (PL). This taxonomy-based survey provides a comprehensive analysis of LLMs in the NL-PL domain, investigating how these models are utilized in coding tasks and examining their methodologies, architectures, and training processes. We propose a taxonomy-based framework that categorizes relevant concepts, providing a unified classification system to facilitate a deeper understanding of this rapidly evolving field. This survey offers insights into the current state and future directions of LLMs in coding tasks, including their applications and limitations.
☆ Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective AAAI 2025
To address catastrophic forgetting in Continual Relation Extraction (CRE), many current approaches rely on memory buffers to rehearse previously learned knowledge while acquiring new tasks. Recently, prompt-based methods have emerged as potent alternatives to rehearsal-based strategies, demonstrating strong empirical performance. However, upon analyzing existing prompt-based approaches for CRE, we identified several critical limitations, such as inaccurate prompt selection, inadequate mechanisms for mitigating forgetting in shared parameters, and suboptimal handling of cross-task and within-task variances. To overcome these challenges, we draw inspiration from the relationship between prefix-tuning and mixture of experts, proposing a novel approach that employs a prompt pool for each task, capturing variations within each task while enhancing cross-task variances. Furthermore, we incorporate a generative model to consolidate prior knowledge within shared parameters, eliminating the need for explicit data storage. Extensive experiments validate the efficacy of our approach, demonstrating superior performance over state-of-the-art prompt-based and rehearsal-free methods in continual relation extraction.
comment: Accepted to AAAI 2025
☆ A Preliminary Analysis of Automatic Word and Syllable Prominence Detection in Non-Native Speech With Text-to-Speech Prosody Embeddings
Automatic detection of prominence at the word and syllable-levels is critical for building computer-assisted language learning systems. It has been shown that prosody embeddings learned by the current state-of-the-art (SOTA) text-to-speech (TTS) systems could generate word- and syllable-level prominence in the synthesized speech as natural as in native speech. To understand the effectiveness of prosody embeddings from TTS for prominence detection under nonnative context, a comparative analysis is conducted on the embeddings extracted from native and non-native speech considering the prominence-related embeddings: duration, energy, and pitch from a SOTA TTS named FastSpeech2. These embeddings are extracted under two conditions considering: 1) only text, 2) both speech and text. For the first condition, the embeddings are extracted directly from the TTS inference mode, whereas for the second condition, we propose to extract from the TTS under training mode. Experiments are conducted on native speech corpus: Tatoeba, and non-native speech corpus: ISLE. For experimentation, word-level prominence locations are manually annotated for both corpora. The highest relative improvement on word \& syllable-level prominence detection accuracies with the TTS embeddings are found to be 13.7% & 5.9% and 16.2% & 6.9% compared to those with the heuristic-based features and self-supervised Wav2Vec-2.0 representations, respectively.
☆ Y-NQ: English-Yorùbá Evaluation dataset for Open-Book Reading Comprehension and Text Generation
The purpose of this work is to share an English-Yor\`ub\'a evaluation dataset for open-book reading comprehension and text generation to assess the performance of models both in a high- and a low- resource language. The dataset contains 358 questions and answers on 338 English documents and 208 Yor\`ub\'a documents. The average document length is ~ 10k words for English and 430 words for Yor\`ub\'a. Experiments show a consistent disparity in performance between the two languages, with Yor\`ub\'a falling behind English for automatic metrics even if documents are much shorter for this language. For a small set of documents with comparable length, performance of Yor\`ub\'a drops by x2.5 times. When analyzing performance by length, we observe that Yor\`ub\'a decreases performance dramatically for documents that reach 1500 words while English performance is barely affected at that length. Our dataset opens the door to showcasing if English LLM reading comprehension capabilities extend to Yor\`ub\'a, which for the evaluated LLMs is not the case.
☆ 2M-BELEBELE: Highly Multilingual Speech and American Sign Language Comprehension Dataset
We introduce the first highly multilingual speech and American Sign Language (ASL) comprehension dataset by extending BELEBELE. Our dataset covers 74 spoken languages at the intersection of BELEBELE and FLEURS, and one sign language (ASL). We evaluate 2M-BELEBELE dataset for both 5-shot and zero-shot settings and across languages, the speech comprehension accuracy is ~ 8% average lower compared to reading comprehension.
☆ LCFO: Long Context and Long Form Output Dataset and Benchmarking
This paper presents the Long Context and Form Output (LCFO) benchmark, a novel evaluation framework for assessing gradual summarization and summary expansion capabilities across diverse domains. LCFO consists of long input documents (5k words average length), each of which comes with three summaries of different lengths (20%, 10%, and 5% of the input text), as well as approximately 15 questions and answers (QA) related to the input content. Notably, LCFO also provides alignments between specific QA pairs and corresponding summaries in 7 domains. The primary motivation behind providing summaries of different lengths is to establish a controllable framework for generating long texts from shorter inputs, i.e. summary expansion. To establish an evaluation metric framework for summarization and summary expansion, we provide human evaluation scores for human-generated outputs, as well as results from various state-of-the-art large language models (LLMs). GPT-4o-mini achieves best human scores among automatic systems in both summarization and summary expansion tasks (~ +10% and +20%, respectively). It even surpasses human output quality in the case of short summaries (~ +7%). Overall automatic metrics achieve low correlations with human evaluation scores (~ 0.4) but moderate correlation on specific evaluation aspects such as fluency and attribution (~ 0.6). The LCFO benchmark offers a standardized platform for evaluating summarization and summary expansion performance, as well as corresponding automatic metrics, thereby providing an important evaluation framework to advance generative AI.
☆ Discrete Subgraph Sampling for Interpretable Graph based Visual Question Answering COLING 2025
Explainable artificial intelligence (XAI) aims to make machine learning models more transparent. While many approaches focus on generating explanations post-hoc, interpretable approaches, which generate the explanations intrinsically alongside the predictions, are relatively rare. In this work, we integrate different discrete subset sampling methods into a graph-based visual question answering system to compare their effectiveness in generating interpretable explanatory subgraphs intrinsically. We evaluate the methods on the GQA dataset and show that the integrated methods effectively mitigate the performance trade-off between interpretability and answer accuracy, while also achieving strong co-occurrences between answer and question tokens. Furthermore, we conduct a human evaluation to assess the interpretability of the generated subgraphs using a comparative setting with the extended Bradley-Terry model, showing that the answer and question token co-occurrence metrics strongly correlate with human preferences. Our source code is publicly available.
comment: Accepted at COLING 2025
☆ Accurate Medical Named Entity Recognition Through Specialized NLP Models
This study evaluated the effect of BioBERT in medical text processing for the task of medical named entity recognition. Through comparative experiments with models such as BERT, ClinicalBERT, SciBERT, and BlueBERT, the results showed that BioBERT achieved the best performance in both precision and F1 score, verifying its applicability and superiority in the medical field. BioBERT enhances its ability to understand professional terms and complex medical texts through pre-training on biomedical data, providing a powerful tool for medical information extraction and clinical decision support. The study also explored the privacy and compliance challenges of BioBERT when processing medical data, and proposed future research directions for combining other medical-specific models to improve generalization and robustness. With the development of deep learning technology, the potential of BioBERT in application fields such as intelligent medicine, personalized treatment, and disease prediction will be further expanded. Future research can focus on the real-time and interpretability of the model to promote its widespread application in the medical field.
☆ TouchTTS: An Embarrassingly Simple TTS Framework that Everyone Can Touch
It is well known that LLM-based systems are data-hungry. Recent LLM-based TTS works typically employ complex data processing pipelines to obtain high-quality training data. These sophisticated pipelines require excellent models at each stage (e.g., speech denoising, speech enhancement, speaker diarization, and punctuation models), which themselves demand high-quality training data and are rarely open-sourced. Even with state-of-the-art models, issues persist, such as incomplete background noise removal and misalignment between punctuation and actual speech pauses. Moreover, the stringent filtering strategies often retain only 10-30\% of the original data, significantly impeding data scaling efforts. In this work, we leverage a noise-robust audio tokenizer (S3Tokenizer) to design a simplified yet effective TTS data processing pipeline that maintains data quality while substantially reducing data acquisition costs, achieving a data retention rate of over 50\%. Beyond data scaling challenges, LLM-based TTS systems also incur higher deployment costs compared to conventional approaches. Current systems typically use LLMs solely for text-to-token generation, while requiring separate models (e.g., flow matching models) for token-to-waveform generation, which cannot be directly executed by LLM inference engines, further complicating deployment. To address these challenges, we eliminate redundant modules in both LLM and flow components, replacing the flow model backbone with an LLM architecture. Building upon this simplified flow backbone, we propose a unified architecture for both streaming and non-streaming inference, significantly reducing deployment costs. Finally, we explore the feasibility of unifying TTS and ASR tasks using the same data for training, thanks to the simplified pipeline and the S3Tokenizer that reduces the quality requirements for TTS training data.
comment: Technical Report
☆ DocSum: Domain-Adaptive Pre-training for Document Abstractive Summarization
Abstractive summarization has made significant strides in condensing and rephrasing large volumes of text into coherent summaries. However, summarizing administrative documents presents unique challenges due to domain-specific terminology, OCR-generated errors, and the scarcity of annotated datasets for model fine-tuning. Existing models often struggle to adapt to the intricate structure and specialized content of such documents. To address these limitations, we introduce DocSum, a domain-adaptive abstractive summarization framework tailored for administrative documents. Leveraging pre-training on OCR-transcribed text and fine-tuning with an innovative integration of question-answer pairs, DocSum enhances summary accuracy and relevance. This approach tackles the complexities inherent in administrative content, ensuring outputs that align with real-world business needs. To evaluate its capabilities, we define a novel downstream task setting-Document Abstractive Summarization-which reflects the practical requirements of business and organizational settings. Comprehensive experiments demonstrate DocSum's effectiveness in producing high-quality summaries, showcasing its potential to improve decision-making and operational workflows across the public and private sectors.
☆ From communities to interpretable network and word embedding: an unified approach
Modelling information from complex systems such as humans social interaction or words co-occurrences in our languages can help to understand how these systems are organized and function. Such systems can be modelled by networks, and network theory provides a useful set of methods to analyze them. Among these methods, graph embedding is a powerful tool to summarize the interactions and topology of a network in a vectorized feature space. When used in input of machine learning algorithms, embedding vectors help with common graph problems such as link prediction, graph matching, etc. Word embedding has the goal of representing the sense of words, extracting it from large text corpora. Despite differences in the structure of information in input of embedding algorithms, many graph embedding approaches are adapted and inspired from methods in NLP. Limits of these methods are observed in both domains. Most of these methods require long and resource greedy training. Another downside to most methods is that they are black-box, from which understanding how the information is structured is rather complex. Interpretability of a model allows understanding how the vector space is structured without the need for external information, and thus can be audited more easily. With both these limitations in mind, we propose a novel framework to efficiently embed network vertices in an interpretable vector space. Our Lower Dimension Bipartite Framework (LDBGF) leverages the bipartite projection of a network using cliques to reduce dimensionality. Along with LDBGF, we introduce two implementations of this framework that rely on communities instead of cliques: SINr-NR and SINr-MF. We show that SINr-MF can perform well on classical graphs and SINr-NR can produce high-quality graph and word embeddings that are interpretable and stable across runs.
☆ Illusory VQA: Benchmarking and Enhancing Multimodal Models on Visual Illusions
In recent years, Visual Question Answering (VQA) has made significant strides, particularly with the advent of multimodal models that integrate vision and language understanding. However, existing VQA datasets often overlook the complexities introduced by image illusions, which pose unique challenges for both human perception and model interpretation. In this study, we introduce a novel task called Illusory VQA, along with four specialized datasets: IllusionMNIST, IllusionFashionMNIST, IllusionAnimals, and IllusionChar. These datasets are designed to evaluate the performance of state-of-the-art multimodal models in recognizing and interpreting visual illusions. We assess the zero-shot performance of various models, fine-tune selected models on our datasets, and propose a simple yet effective solution for illusion detection using Gaussian and blur low-pass filters. We show that this method increases the performance of models significantly and in the case of BLIP-2 on IllusionAnimals without any fine-tuning, it outperforms humans. Our findings highlight the disparity between human and model perception of illusions and demonstrate that fine-tuning and specific preprocessing techniques can significantly enhance model robustness. This work contributes to the development of more human-like visual understanding in multimodal models and suggests future directions for adapting filters using learnable parameters.
☆ NLPineers@ NLU of Devanagari Script Languages 2025: Hate Speech Detection using Ensembling of BERT-based models
This paper explores hate speech detection in Devanagari-scripted languages, focusing on Hindi and Nepali, for Subtask B of the CHIPSAL@COLING 2025 Shared Task. Using a range of transformer-based models such as XLM-RoBERTa, MURIL, and IndicBERT, we examine their effectiveness in navigating the nuanced boundary between hate speech and free expression. Our best performing model, implemented as ensemble of multilingual BERT models achieve Recall of 0.7762 (Rank 3/31 in terms of recall) and F1 score of 0.6914 (Rank 17/31). To address class imbalance, we used backtranslation for data augmentation, and cosine similarity to preserve label consistency after augmentation. This work emphasizes the need for hate speech detection in Devanagari-scripted languages and presents a foundation for further research.
☆ How Vision-Language Tasks Benefit from Large Pre-trained Models: A Survey
The exploration of various vision-language tasks, such as visual captioning, visual question answering, and visual commonsense reasoning, is an important area in artificial intelligence and continuously attracts the research community's attention. Despite the improvements in overall performance, classic challenges still exist in vision-language tasks and hinder the development of this area. In recent years, the rise of pre-trained models is driving the research on vision-language tasks. Thanks to the massive scale of training data and model parameters, pre-trained models have exhibited excellent performance in numerous downstream tasks. Inspired by the powerful capabilities of pre-trained models, new paradigms have emerged to solve the classic challenges. Such methods have become mainstream in current research with increasing attention and rapid advances. In this paper, we present a comprehensive overview of how vision-language tasks benefit from pre-trained models. First, we review several main challenges in vision-language tasks and discuss the limitations of previous solutions before the era of pre-training. Next, we summarize the recent advances in incorporating pre-trained models to address the challenges in vision-language tasks. Finally, we analyze the potential risks associated with the inherent limitations of pre-trained models and discuss possible solutions, attempting to provide future research directions.
comment: Under Review
☆ Evil twins are not that evil: Qualitative insights into machine-generated prompts
It has been widely observed that language models (LMs) respond in predictable ways to algorithmically generated prompts that are seemingly unintelligible. This is both a sign that we lack a full understanding of how LMs work, and a practical challenge, because opaqueness can be exploited for harmful uses of LMs, such as jailbreaking. We present the first thorough analysis of opaque machine-generated prompts, or autoprompts, pertaining to 3 LMs of different sizes and families. We find that machine-generated prompts are characterized by a last token that is often intelligible and strongly affects the generation. A small but consistent proportion of the previous tokens are fillers that probably appear in the prompt as a by-product of the fact that the optimization process fixes the number of tokens. The remaining tokens tend to have at least a loose semantic relation with the generation, although they do not engage in well-formed syntactic relations with it. We find moreover that some of the ablations we applied to machine-generated prompts can also be applied to natural language sequences, leading to similar behavior, suggesting that autoprompts are a direct consequence of the way in which LMs process linguistic inputs in general.
☆ Progressive Multi-granular Alignments for Grounded Reasoning in Large Vision-Language Models
Existing Large Vision-Language Models (LVLMs) excel at matching concepts across multi-modal inputs but struggle with compositional concepts and high-level relationships between entities. This paper introduces Progressive multi-granular Vision-Language alignments (PromViL), a novel framework to enhance LVLMs' ability in performing grounded compositional visual reasoning tasks. Our approach constructs a hierarchical structure of multi-modal alignments, ranging from simple to complex concepts. By progressively aligning textual descriptions with corresponding visual regions, our model learns to leverage contextual information from lower levels to inform higher-level reasoning. To facilitate this learning process, we introduce a data generation process that creates a novel dataset derived from Visual Genome, providing a wide range of nested compositional vision-language pairs. Experimental results demonstrate that our PromViL framework significantly outperforms baselines on various visual grounding and compositional question answering tasks.
☆ LatentSpeech: Latent Diffusion for Text-To-Speech Generation
Diffusion-based Generative AI gains significant attention for its superior performance over other generative techniques like Generative Adversarial Networks and Variational Autoencoders. While it has achieved notable advancements in fields such as computer vision and natural language processing, their application in speech generation remains under-explored. Mainstream Text-to-Speech systems primarily map outputs to Mel-Spectrograms in the spectral space, leading to high computational loads due to the sparsity of MelSpecs. To address these limitations, we propose LatentSpeech, a novel TTS generation approach utilizing latent diffusion models. By using latent embeddings as the intermediate representation, LatentSpeech reduces the target dimension to 5% of what is required for MelSpecs, simplifying the processing for the TTS encoder and vocoder and enabling efficient high-quality speech generation. This study marks the first integration of latent diffusion models in TTS, enhancing the accuracy and naturalness of generated speech. Experimental results on benchmark datasets demonstrate that LatentSpeech achieves a 25% improvement in Word Error Rate and a 24% improvement in Mel Cepstral Distortion compared to existing models, with further improvements rising to 49.5% and 26%, respectively, with additional training data. These findings highlight the potential of LatentSpeech to advance the state-of-the-art in TTS technology
☆ Aligner-Guided Training Paradigm: Advancing Text-to-Speech Models with Aligner Guided Duration
Recent advancements in text-to-speech (TTS) systems, such as FastSpeech and StyleSpeech, have significantly improved speech generation quality. However, these models often rely on duration generated by external tools like the Montreal Forced Aligner, which can be time-consuming and lack flexibility. The importance of accurate duration is often underestimated, despite their crucial role in achieving natural prosody and intelligibility. To address these limitations, we propose a novel Aligner-Guided Training Paradigm that prioritizes accurate duration labelling by training an aligner before the TTS model. This approach reduces dependence on external tools and enhances alignment accuracy. We further explore the impact of different acoustic features, including Mel-Spectrograms, MFCCs, and latent features, on TTS model performance. Our experimental results show that aligner-guided duration labelling can achieve up to a 16\% improvement in word error rate and significantly enhance phoneme and tone alignment. These findings highlight the effectiveness of our approach in optimizing TTS systems for more natural and intelligible speech generation.
☆ Seeing Syntax: Uncovering Syntactic Learning Limitations in Vision-Language Models
Vision-language models (VLMs), serve as foundation models for multi-modal applications such as image captioning and text-to-image generation. Recent studies have highlighted limitations in VLM text encoders, particularly in areas like compositionality and semantic understanding, though the underlying reasons for these limitations remain unclear. In this work, we aim to address this gap by analyzing the syntactic information, one of the fundamental linguistic properties, encoded by the text encoders of VLMs. We perform a thorough analysis comparing VLMs with different objective functions, parameter size and training data size, and with uni-modal language models (ULMs) in their ability to encode syntactic knowledge. Our findings suggest that ULM text encoders acquire syntactic information more effectively than those in VLMs. The syntactic information learned by VLM text encoders is shaped primarily by the pre-training objective, which plays a more crucial role than other factors such as model architecture, model size, or the volume of pre-training data. Models exhibit different layer-wise trends where CLIP performance dropped across layers while for other models, middle layers are rich in encoding syntactic knowledge.
☆ Barking Up The Syntactic Tree: Enhancing VLM Training with Syntactic Losses
Vision-Language Models (VLMs) achieved strong performance on a variety of tasks (e.g., image-text retrieval, visual question answering). However, most VLMs rely on coarse-grained image-caption pairs for alignment, relying on data volume to resolve ambiguities and ground linguistic concepts in images. The richer semantic and syntactic structure within text is largely overlooked. To address this, we propose HIerarchically STructured Learning (HIST) that enhances VLM training without any additional supervision, by hierarchically decomposing captions into the constituent Subject, Noun Phrases, and Composite Phrases. Entailment between these constituent components allows us to formulate additional regularization constraints on the VLM attention maps. Specifically, we introduce two novel loss functions: (1) Subject Loss, which aligns image content with the subject of corresponding phrase, acting as an entailment of standard contrastive/matching losses at the Phrase level; (2) Addition Loss, to balance attention across multiple objects. HIST is general, and can be applied to any VLM for which attention between vision and language can be computed; we illustrate its efficacy on BLIP and ALBEF. HIST outperforms baseline VLMs, achieving up to +9.8% improvement in visual grounding, +6.3% in multi-object referring segmentation, +1.1% in image-text retrieval, and +0.2% in visual question answering, underscoring the value of structuring learning in VLMs.
☆ Doubly-Universal Adversarial Perturbations: Deceiving Vision-Language Models Across Both Images and Text with a Single Perturbation
Large Vision-Language Models (VLMs) have demonstrated remarkable performance across multimodal tasks by integrating vision encoders with large language models (LLMs). However, these models remain vulnerable to adversarial attacks. Among such attacks, Universal Adversarial Perturbations (UAPs) are especially powerful, as a single optimized perturbation can mislead the model across various input images. In this work, we introduce a novel UAP specifically designed for VLMs: the Doubly-Universal Adversarial Perturbation (Doubly-UAP), capable of universally deceiving VLMs across both image and text inputs. To successfully disrupt the vision encoder's fundamental process, we analyze the core components of the attention mechanism. After identifying value vectors in the middle-to-late layers as the most vulnerable, we optimize Doubly-UAP in a label-free manner with a frozen model. Despite being developed as a black-box to the LLM, Doubly-UAP achieves high attack success rates on VLMs, consistently outperforming baseline methods across vision-language tasks. Extensive ablation studies and analyses further demonstrate the robustness of Doubly-UAP and provide insights into how it influences internal attention mechanisms.
☆ Adversarial Vulnerabilities in Large Language Models for Time Series Forecasting
Large Language Models (LLMs) have recently demonstrated significant potential in the field of time series forecasting, offering impressive capabilities in handling complex temporal data. However, their robustness and reliability in real-world applications remain under-explored, particularly concerning their susceptibility to adversarial attacks. In this paper, we introduce a targeted adversarial attack framework for LLM-based time series forecasting. By employing both gradient-free and black-box optimization methods, we generate minimal yet highly effective perturbations that significantly degrade the forecasting accuracy across multiple datasets and LLM architectures. Our experiments, which include models like TimeGPT and LLM-Time with GPT-3.5, GPT-4, LLaMa, and Mistral, show that adversarial attacks lead to much more severe performance degradation than random noise, and demonstrate the broad effectiveness of our attacks across different LLMs. The results underscore the critical vulnerabilities of LLMs in time series forecasting, highlighting the need for robust defense mechanisms to ensure their reliable deployment in practical applications.
comment: 11 pages, 5 figures
☆ Multilingual LLMs Inherently Reward In-Language Time-Sensitive Semantic Alignment for Low-Resource Languages
The unwavering disparity in labeled resources between resource-rich languages and those considered low-resource remains a significant impediment for Large Language Models (LLMs). Recent strides in cross-lingual in-context learning (X-ICL), mainly through semantically aligned examples retrieved from multilingual pre-trained transformers, have shown promise in mitigating this issue. However, our investigation reveals that LLMs intrinsically reward in-language semantically aligned cross-lingual instances over direct cross-lingual semantic alignments, with a pronounced disparity in handling time-sensitive queries in the X-ICL setup. Such queries demand sound temporal reasoning ability from LLMs, yet the advancements have predominantly focused on English. This study aims to bridge this gap by improving temporal reasoning capabilities in low-resource languages. To this end, we introduce mTEMPREASON a temporal reasoning dataset aimed at the varied degrees of low-resource languages and propose Cross-Lingual Time-Sensitive Semantic Alignment (CLiTSSA), a novel method to improve temporal reasoning in these contexts. To facilitate this, we construct an extension of mTEMPREASON comprising pairs of parallel cross-language temporal queries along with their anticipated in-language semantic similarity scores. Our empirical evidence underscores the superior performance of CLiTSSA compared to established baselines across three languages - Romanian, German, and French, encompassing three temporal tasks and including a diverse set of four contemporaneous LLMs. This marks a significant step forward in addressing resource disparity in the context of temporal reasoning across languages.
☆ Federated In-Context LLM Agent Learning
Large Language Models (LLMs) have revolutionized intelligent services by enabling logical reasoning, tool use, and interaction with external systems as agents. The advancement of LLMs is frequently hindered by the scarcity of high-quality data, much of which is inherently sensitive. Federated learning (FL) offers a potential solution by facilitating the collaborative training of distributed LLMs while safeguarding private data. However, FL frameworks face significant bandwidth and computational demands, along with challenges from heterogeneous data distributions. The emerging in-context learning capability of LLMs offers a promising approach by aggregating natural language rather than bulky model parameters. Yet, this method risks privacy leakage, as it necessitates the collection and presentation of data samples from various clients during aggregation. In this paper, we propose a novel privacy-preserving Federated In-Context LLM Agent Learning (FICAL) algorithm, which to our best knowledge for the first work unleashes the power of in-context learning to train diverse LLM agents through FL. In our design, knowledge compendiums generated by a novel LLM-enhanced Knowledge Compendiums Generation (KCG) module are transmitted between clients and the server instead of model parameters in previous FL methods. Apart from that, an incredible Retrieval Augmented Generation (RAG) based Tool Learning and Utilizing (TLU) module is designed and we incorporate the aggregated global knowledge compendium as a teacher to teach LLM agents the usage of tools. We conducted extensive experiments and the results show that FICAL has competitive performance compared to other SOTA baselines with a significant communication cost decrease of $\mathbf{3.33\times10^5}$ times.
☆ M2SE: A Multistage Multitask Instruction Tuning Strategy for Unified Sentiment and Emotion Analysis
Sentiment analysis and emotion recognition are crucial for applications such as human-computer interaction and depression detection. Traditional unimodal methods often fail to capture the complexity of emotional expressions due to conflicting signals from different modalities. Current Multimodal Large Language Models (MLLMs) also face challenges in detecting subtle facial expressions and addressing a wide range of emotion-related tasks. To tackle these issues, we propose M2SE, a Multistage Multitask Sentiment and Emotion Instruction Tuning Strategy for general-purpose MLLMs. It employs a combined approach to train models on tasks such as multimodal sentiment analysis, emotion recognition, facial expression recognition, emotion reason inference, and emotion cause-pair extraction. We also introduce the Emotion Multitask dataset (EMT), a custom dataset that supports these five tasks. Our model, Emotion Universe (EmoVerse), is built on a basic MLLM framework without modifications, yet it achieves substantial improvements across these tasks when trained with the M2SE strategy. Extensive experiments demonstrate that EmoVerse outperforms existing methods, achieving state-of-the-art results in sentiment and emotion tasks. These results highlight the effectiveness of M2SE in enhancing multimodal emotion perception. The dataset and code are available at https://github.com/xiaoyaoxinyi/M2SE.
☆ Bootstrapping Heterogeneous Graph Representation Learning via Large Language Models: A Generalized Approach AAAI 2025
Graph representation learning methods are highly effective in handling complex non-Euclidean data by capturing intricate relationships and features within graph structures. However, traditional methods face challenges when dealing with heterogeneous graphs that contain various types of nodes and edges due to the diverse sources and complex nature of the data. Existing Heterogeneous Graph Neural Networks (HGNNs) have shown promising results but require prior knowledge of node and edge types and unified node feature formats, which limits their applicability. Recent advancements in graph representation learning using Large Language Models (LLMs) offer new solutions by integrating LLMs' data processing capabilities, enabling the alignment of various graph representations. Nevertheless, these methods often overlook heterogeneous graph data and require extensive preprocessing. To address these limitations, we propose a novel method that leverages the strengths of both LLM and GNN, allowing for the processing of graph data with any format and type of nodes and edges without the need for type information or special preprocessing. Our method employs LLM to automatically summarize and classify different data formats and types, aligns node features, and uses a specialized GNN for targeted learning, thus obtaining effective graph representations for downstream tasks. Theoretical analysis and experimental validation have demonstrated the effectiveness of our method.
comment: Accepted by AAAI 2025
☆ TinyThinker: Distilling Reasoning through Coarse-to-Fine Knowledge Internalization with Self-Reflection
Large Language Models exhibit impressive reasoning capabilities across diverse tasks, motivating efforts to distill these capabilities into smaller models through generated reasoning data. However, direct training on such synthesized reasoning data may lead to superficial imitation of reasoning process, rather than fostering a genuine integration of reasoning capabilities with underlying knowledge. To address this, we propose TinyThinker, a framework introducing two novel approaches. First, we introduce a three-stage process that incrementally guides the student model through the reasoning process, progressively refining knowledge from coarse to fine granularity. Second, we develop a two-phase training framework comprising an initial reasoning acquisition phase followed by a self-reflection phase utilizing self-generated data. Experiments on commonsense reasoning benchmarks demonstrate that TinyThinker achieves superior performance compared to baselines. Ablation studies further validate the effectiveness of each component in our framework. TinyThinker is extendable to other knowledge-intensive reasoning tasks, offering an alternative strategy for developing effective reasoning capabilities in smaller language models. Codes are available at https://github.com/shengminp/TinyThinker
☆ Concept Bottleneck Large Language Models
We introduce the Concept Bottleneck Large Language Model (CB-LLM), a pioneering approach to creating inherently interpretable Large Language Models (LLMs). Unlike traditional black-box LLMs that rely on post-hoc interpretation methods with limited neuron function insights, CB-LLM sets a new standard with its built-in interpretability, scalability, and ability to provide clear, accurate explanations. We investigate two essential tasks in the NLP domain: text classification and text generation. In text classification, CB-LLM narrows the performance gap with traditional black-box models and provides clear interpretability. In text generation, we show how interpretable neurons in CB-LLM can be used for concept detection and steering text generation. Our CB-LLMs enable greater interaction between humans and LLMs across a variety of tasks -- a feature notably absent in existing LLMs. Our code is available at https://github.com/Trustworthy-ML-Lab/CB-LLMs.
comment: arXiv admin note: substantial text overlap with arXiv:2407.04307
♻ ☆ Fusing Domain-Specific Content from Large Language Models into Knowledge Graphs for Enhanced Zero Shot Object State Classification AAAI
Domain-specific knowledge can significantly contribute to addressing a wide variety of vision tasks. However, the generation of such knowledge entails considerable human labor and time costs. This study investigates the potential of Large Language Models (LLMs) in generating and providing domain-specific information through semantic embeddings. To achieve this, an LLM is integrated into a pipeline that utilizes Knowledge Graphs and pre-trained semantic vectors in the context of the Vision-based Zero-shot Object State Classification task. We thoroughly examine the behavior of the LLM through an extensive ablation study. Our findings reveal that the integration of LLM-based embeddings, in combination with general-purpose pre-trained embeddings, leads to substantial performance improvements. Drawing insights from this ablation study, we conduct a comparative analysis against competing models, thereby highlighting the state-of-the-art performance achieved by the proposed approach.
comment: Accepted at the AAAI-MAKE 2024
♻ ☆ Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Prompts
As large language models (LLMs) become increasingly prevalent across many real-world applications, understanding and enhancing their robustness to adversarial attacks is of paramount importance. Existing methods for identifying adversarial prompts tend to focus on specific domains, lack diversity, or require extensive human annotations. To address these limitations, we present Rainbow Teaming, a novel black-box approach for producing a diverse collection of adversarial prompts. Rainbow Teaming casts adversarial prompt generation as a quality-diversity problem and uses open-ended search to generate prompts that are both effective and diverse. Focusing on the safety domain, we use Rainbow Teaming to target various state-of-the-art LLMs, including the Llama 2 and Llama 3 models. Our approach reveals hundreds of effective adversarial prompts, with an attack success rate exceeding 90% across all tested models. Furthermore, we demonstrate that prompts generated by Rainbow Teaming are highly transferable and that fine-tuning models with synthetic data generated by our method significantly enhances their safety without sacrificing general performance or helpfulness. We additionally explore the versatility of Rainbow Teaming by applying it to question answering and cybersecurity, showcasing its potential to drive robust open-ended self-improvement in a wide range of applications.
♻ ☆ Comparative Analysis of Pooling Mechanisms in LLMs: A Sentiment Analysis Perspective
Large Language Models (LLMs) have revolutionized natural language processing (NLP) by delivering state-of-the-art performance across a variety of tasks. Among these, Transformer-based models like BERT and GPT rely on pooling layers to aggregate token-level embeddings into sentence-level representations. Common pooling mechanisms such as Mean, Max, and Weighted Sum play a pivotal role in this aggregation process. Despite their widespread use, the comparative performance of these strategies on different LLM architectures remains underexplored. To address this gap, this paper investigates the effects of these pooling mechanisms on two prominent LLM families -- BERT and GPT, in the context of sentence-level sentiment analysis. Comprehensive experiments reveal that each pooling mechanism exhibits unique strengths and weaknesses depending on the task's specific requirements. Our findings underline the importance of selecting pooling methods tailored to the demands of particular applications, prompting a re-evaluation of common assumptions regarding pooling operations. By offering actionable insights, this study contributes to the optimization of LLM-based models for downstream tasks.
comment: 4 figures
♻ ☆ Vision-and-Language Pretraining
With the burgeoning amount of data of image-text pairs and diversity of Vision-and-Language (V\&L) tasks, scholars have introduced an abundance of deep learning models in this research domain. Furthermore, in recent years, transfer learning has also shown tremendous success in Computer Vision for tasks such as Image Classification, Object Detection, etc., and in Natural Language Processing for Question Answering, Machine Translation, etc. Inheriting the spirit of Transfer Learning, research works in V\&L have devised multiple pretraining techniques on large-scale datasets in order to enhance the performance of downstream tasks. The aim of this article is to provide a comprehensive revision of contemporary V\&L pretraining models. In particular, we categorize and delineate pretraining approaches, along with the summary of state-of-the-art vision-and-language pretrained models. Moreover, a list of training datasets and downstream tasks is supplied to further polish the perspective into V\&L pretraining. Lastly, we decided to take a further step to discuss numerous directions for future research.
comment: The content of the paper has been outdated. I would like to rewrite a new version with completely new information.
♻ ☆ VoiceBench: Benchmarking LLM-Based Voice Assistants
Building on the success of large language models (LLMs), recent advancements such as GPT-4o have enabled real-time speech interactions through LLM-based voice assistants, offering a significantly improved user experience compared to traditional text-based interactions. However, the absence of benchmarks designed to evaluate these speech interaction capabilities has hindered progress of LLM-based voice assistants development. Current evaluations focus primarily on automatic speech recognition (ASR) or general knowledge evaluation with clean speeches, neglecting the more intricate, real-world scenarios that involve diverse speaker characteristics, environmental and content factors. To address this, we introduce VoiceBench, the first benchmark designed to provide a multi-faceted evaluation of LLM-based voice assistants. VoiceBench also includes both real and synthetic spoken instructions that incorporate the above three key real-world variations. Extensive experiments reveal the limitations of current LLM-based voice assistant models and offer valuable insights for future research and development in this field.
comment: Work in progress. Data is available at https://github.com/MatthewCYM/VoiceBench
♻ ☆ Topic Classification of Case Law Using a Large Language Model and a New Taxonomy for UK Law: AI Insights into Summary Judgment
This paper addresses a critical gap in legal analytics by developing and applying a novel taxonomy for topic classification of summary judgment cases in the United Kingdom. Using a curated dataset of summary judgment cases, we use the Large Language Model Claude 3 Opus to explore functional topics and trends. We find that Claude 3 Opus correctly classified the topic with an accuracy of 87.13% and an F1 score of 0.87. The analysis reveals distinct patterns in the application of summary judgments across various legal domains. As case law in the United Kingdom is not originally labelled with keywords or a topic filtering option, the findings not only refine our understanding of the thematic underpinnings of summary judgments but also illustrate the potential of combining traditional and AI-driven approaches in legal classification. Therefore, this paper provides a new and general taxonomy for UK law. The implications of this work serve as a foundation for further research and policy discussions in the field of judicial administration and computational legal research methodologies.
♻ ☆ MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering
We introduce MLE-bench, a benchmark for measuring how well AI agents perform at machine learning engineering. To this end, we curate 75 ML engineering-related competitions from Kaggle, creating a diverse set of challenging tasks that test real-world ML engineering skills such as training models, preparing datasets, and running experiments. We establish human baselines for each competition using Kaggle's publicly available leaderboards. We use open-source agent scaffolds to evaluate several frontier language models on our benchmark, finding that the best-performing setup--OpenAI's o1-preview with AIDE scaffolding--achieves at least the level of a Kaggle bronze medal in 16.9% of competitions. In addition to our main results, we investigate various forms of resource scaling for AI agents and the impact of contamination from pre-training. We open-source our benchmark code (github.com/openai/mle-bench/) to facilitate future research in understanding the ML engineering capabilities of AI agents.
comment: 10 pages, 17 pages appendix. Equal contribution by first seven authors, authors randomized. Corrected footnote 4
♻ ☆ Filipino Benchmarks for Measuring Sexist and Homophobic Bias in Multilingual Language Models from Southeast Asia COLING 2025
Bias studies on multilingual models confirm the presence of gender-related stereotypes in masked models processing languages with high NLP resources. We expand on this line of research by introducing Filipino CrowS-Pairs and Filipino WinoQueer: benchmarks that assess both sexist and anti-queer biases in pretrained language models (PLMs) handling texts in Filipino, a low-resource language from the Philippines. The benchmarks consist of 7,074 new challenge pairs resulting from our cultural adaptation of English bias evaluation datasets, a process that we document in detail to guide similar forthcoming efforts. We apply the Filipino benchmarks on masked and causal multilingual models, including those pretrained on Southeast Asian data, and find that they contain considerable amounts of bias. We also find that for multilingual models, the extent of bias learned for a particular language is influenced by how much pretraining data in that language a model was exposed to. Our benchmarks and insights can serve as a foundation for future work analyzing and mitigating bias in multilingual models.
comment: Accepted for presentation at The First Workshop on Language Models for Low-Resource Languages (LoResLM) at The 31st International Conference on Computational Linguistics (COLING 2025)
♻ ☆ Understanding Token Probability Encoding in Output Embeddings COLING 2025
In this paper, we investigate the output token probability information in the output embedding of language models. We find an approximate common log-linear encoding of output token probabilities within the output embedding vectors and empirically demonstrate that it is accurate and sparse. As a causality examination, we steer the encoding in output embedding to modify the output probability distribution accurately. Moreover, the sparsity we find in output probability encoding suggests that a large number of dimensions in the output embedding do not contribute to causal language modeling. Therefore, we attempt to delete the output-unrelated dimensions and find more than 30% of the dimensions can be deleted without significant movement in output distribution and sequence generation. Additionally, in the pre-training dynamics of language models, we find that the output embeddings capture the corpus token frequency information in early steps, even before an obvious convergence of parameters starts.
comment: 15 pages, 17 figures, 3 tables. COLING 2025 Accepted
♻ ☆ Searching for Structure: Investigating Emergent Communication with Large Language Models
Human languages have evolved to be structured through repeated language learning and use. These processes introduce biases that operate during language acquisition and shape linguistic systems toward communicative efficiency. In this paper, we investigate whether the same happens if artificial languages are optimised for implicit biases of Large Language Models (LLMs). To this end, we simulate a classical referential game in which LLMs learn and use artificial languages. Our results show that initially unstructured holistic languages are indeed shaped to have some structural properties that allow two LLM agents to communicate successfully. Similar to observations in human experiments, generational transmission increases the learnability of languages, but can at the same time result in non-humanlike degenerate vocabularies. Taken together, this work extends experimental findings, shows that LLMs can be used as tools in simulations of language evolution, and opens possibilities for future human-machine experiments in this field.
♻ ☆ How to Leverage Digit Embeddings to Represent Numbers? COLING 2025
Within numerical reasoning, understanding numbers themselves is still a challenge for existing language models. Simple generalisations, such as solving 100+200 instead of 1+2, can substantially affect model performance (Sivakumar and Moosavi, 2023). Among various techniques, character-level embeddings of numbers have emerged as a promising approach to improve number representation. However, this method has limitations as it leaves the task of aggregating digit representations to the model, which lacks direct supervision for this process. In this paper, we explore the use of mathematical priors to compute aggregated digit embeddings and explicitly incorporate these aggregates into transformer models. This can be achieved either by adding a special token to the input embeddings or by introducing an additional loss function to enhance correct predictions. We evaluate the effectiveness of incorporating this explicit aggregation, analysing its strengths and shortcomings, and discuss future directions to better benefit from this approach. Our methods, while simple, are compatible with any pretrained model, easy to implement, and have been made publicly available.
comment: Accepted at COLING 2025
♻ ☆ Semantic Loss Guided Data Efficient Supervised Fine Tuning for Safe Responses in LLMs
Large Language Models (LLMs) generating unsafe responses to toxic prompts is a significant issue in their applications. While various efforts aim to address this safety concern, previous approaches often demand substantial human data collection or rely on the less dependable option of using another LLM to generate corrective data. In this paper, we aim to take this problem and overcome limitations of requiring significant high-quality human data. Our method requires only a small set of unsafe responses to toxic prompts, easily obtained from the unsafe LLM itself. By employing a semantic cost combined with a negative Earth Mover Distance (EMD) loss, we guide the LLM away from generating unsafe responses. Additionally, we propose a novel lower bound for EMD loss, enabling more efficient optimization. Our results demonstrate superior performance and data efficiency compared to baselines, and we further examine the nuanced effects of over-alignment and potential degradation of language capabilities when using contrastive data.
♻ ☆ HonestLLM: Toward an Honest and Helpful Large Language Model
Large Language Models (LLMs) have achieved remarkable success across various industries due to their exceptional generative capabilities. However, for safe and effective real-world deployments, ensuring honesty and helpfulness is critical. This paper addresses the question: Can we prioritize the helpfulness of LLMs while preserving their honesty? To begin with, we establish exhaustive principles aimed at guaranteeing the honesty of LLM. Additionally, we introduce a novel dataset, referred to as HoneSet, comprising 930 queries spanning six categories meticulously crafted to assess an LLM's capacity for maintaining honesty. Subsequently, we present two approaches to augmenting honesty and helpfulness in LLMs: a training-free enhancement and a fine-tuning-based improvement. The training-free approach, which is based on curiosity-driven prompting, empowers LLMs to articulate internal confusion and uncertainty regarding queries, thereby optimizing their responses. Conversely, the fine-tuning-based method employs a two-stage process inspired by curriculum learning: initially instructing LLMs to discern between honest and dishonest responses, then refining their training to enhance helpfulness. Experiments conducted on nine prominent LLMs demonstrate a significant improvement in alignment with honesty across all models through the implementation of our proposed enhancements. Particularly noteworthy is the 65.3% enhancement observed in Llama3-8b and the remarkable 124.7% improvement in Mistral-7b, as measured by the H$^{2}$ (honest and helpful) assessment. We believe that our work can pave the way for developing more trustworthy LLMs for real-world applications.
♻ ☆ LA4SR: illuminating the dark proteome with generative AI
AI language models (LMs) show promise for biological sequence analysis. We re-engineered open-source LMs (GPT-2, BLOOM, DistilRoBERTa, ELECTRA, and Mamba, ranging from 70M to 12B parameters) for microbial sequence classification. The models achieved F1 scores up to 95 and operated 16,580x faster and at 2.9x the recall of BLASTP. They effectively classified the algal dark proteome - uncharacterized proteins comprising about 65% of total proteins - validated on new data including a new, complete Hi-C/Pacbio Chlamydomonas genome. Larger (>1B) LA4SR models reached high accuracy (F1 > 86) when trained on less than 2% of available data, rapidly achieving strong generalization capacity. High accuracy was achieved when training data had intact or scrambled terminal information, demonstrating robust generalization to incomplete sequences. Finally, we provide custom AI explainability software tools for attributing amino acid patterns to AI generative processes and interpret their outputs in evolutionary and biophysical contexts.
♻ ☆ FTFT: Efficient and Robust Fine-Tuning by Transferring Training Dynamics COLING 2025
Despite the massive success of fine-tuning Pre-trained Language Models (PLMs), they remain susceptible to out-of-distribution input. Dataset cartography is a simple yet effective dual-model approach that improves the robustness of fine-tuned PLMs. It involves fine-tuning a model on the original training set (i.e. reference model), selecting a subset of important training instances based on the training dynamics, and fine-tuning again only on these selected examples (i.e. main model). However, this approach requires fine-tuning the same model twice, which is computationally expensive for large PLMs. In this paper, we show that (1) training dynamics are highly transferable across model sizes and pre-training methods, and that (2) fine-tuning main models using these selected training instances achieves higher training efficiency than empirical risk minimization (ERM). Building on these observations, we propose a novel fine-tuning approach: Fine-Tuning by transFerring Training dynamics (FTFT). Compared with dataset cartography, FTFT uses more efficient reference models and aggressive early stopping. FTFT achieves robustness improvements over ERM while lowering the training cost by up to $\sim 50\%$.
comment: COLING 2025 Camera-Ready
♻ ☆ One Token Can Help! Learning Scalable and Pluggable Virtual Tokens for Retrieval-Augmented Large Language Models AAAI 2025
Retrieval-augmented generation (RAG) is a promising way to improve large language models (LLMs) for generating more factual, accurate, and up-to-date content. Existing methods either optimize prompts to guide LLMs in leveraging retrieved information or directly fine-tune LLMs to adapt to RAG scenarios. Although fine-tuning can yield better performance, it often compromises the LLMs' general generation capabilities by modifying their parameters. This limitation poses challenges in practical applications, especially when LLMs are already deployed, as parameter adjustments may affect their original functionality. To address this, we propose a novel method that involves learning scalable and pluggable virtual tokens for RAG. By maintaining the LLMs' original parameters and fine-tuning only the embeddings of these pluggable tokens, our approach not only enhances LLMs' performance but also preserves their general generation capabilities. Furthermore, we design several training strategies to improve the scalability, flexibility, and generalizability of our method. Comprehensive experiments across 12 question-answering tasks demonstrate the superiority of our approach.
comment: Accepted by AAAI 2025, repo: https://github.com/DaoD/SPRING/
♻ ☆ Improving Mortality Prediction After Radiotherapy with Large Language Model Structuring of Large-Scale Unstructured Electronic Health Records
Accurate survival prediction in radiotherapy (RT) is critical for optimizing treatment decisions. This study developed and validated the RT-Surv framework, which integrates general-domain, open-source large language models (LLMs) to structure unstructured electronic health records alongside structured clinical data. Using data from 34,276 patients and an external cohort of 852, the framework successfully transformed unstructured clinical information into structured formats. Incorporating LLM-structured clinical features improved the concordance index from 0.779 to 0.842 during external validation, demonstrating a significant performance enhancement. Key LLM-structured features, such as disease extent, general condition, and RT purpose, showed high predictive importance and aligned closely with statistically significant predictors identified through conventional statistical analyses, thereby improving model interpretability. Furthermore, the framework enhanced risk stratification, enabling more distinct differentiation among low-, intermediate-, and high-risk groups (p < 0.001) using LLM-structured clinical features. These findings highlight the potential of LLMs to convert unstructured data into actionable insights, improving predictive modeling and patient outcomes in clinics.
comment: 23 pages, 2 tables, 4 figures
♻ ☆ GrEmLIn: A Repository of Green Baseline Embeddings for 87 Low-Resource Languages Injected with Multilingual Graph Knowledge
Contextualized embeddings based on large language models (LLMs) are available for various languages, but their coverage is often limited for lower resourced languages. Using LLMs for such languages is often difficult due to a high computational cost; not only during training, but also during inference. Static word embeddings are much more resource-efficient ("green"), and thus still provide value, particularly for very low-resource languages. There is, however, a notable lack of comprehensive repositories with such embeddings for diverse languages. To address this gap, we present GrEmLIn, a centralized repository of green, static baseline embeddings for 87 mid- and low-resource languages. We compute GrEmLIn embeddings with a novel method that enhances GloVe embeddings by integrating multilingual graph knowledge, which makes our static embeddings competitive with LLM representations, while being parameter-free at inference time. Our experiments demonstrate that GrEmLIn embeddings outperform state-of-the-art contextualized embeddings from E5 on the task of lexical similarity. They remain competitive in extrinsic evaluation tasks like sentiment analysis and natural language inference, with average performance gaps of just 5-10\% or less compared to state-of-the-art models, given a sufficient vocabulary overlap with the target task, and underperform only on topic classification. Our code and embeddings are publicly available at https://huggingface.co/DFKI.
comment: Long paper, preview
♻ ☆ Using Game Play to Investigate Multimodal and Conversational Grounding in Large Multimodal Models COLING 2025
While the situation has improved for text-only models, it again seems to be the case currently that multimodal (text and image) models develop faster than ways to evaluate them. In this paper, we bring a recently developed evaluation paradigm from text models to multimodal models, namely evaluation through the goal-oriented game (self) play, complementing reference-based and preference-based evaluation. Specifically, we define games that challenge a model's capability to represent a situation from visual information and align such representations through dialogue. We find that the largest closed models perform rather well on the games that we define, while even the best open-weight models struggle with them. On further analysis, we find that the exceptional deep captioning capabilities of the largest models drive some of the performance. There is still room to grow for both kinds of models, ensuring the continued relevance of the benchmark.
comment: Accepted at COLING 2025
♻ ☆ CNNSum: Exploring Long-Context Summarization with Large Language Models in Chinese Novels
Large Language Models (LLMs) have been well-researched in many long-context tasks. However, due to high annotation costs, high-quality long-context summary datasets for training or evaluation are scarce, limiting further research. In this work, we introduce CNNSum, a new multi-scale Chinese long-context novel summarization benchmark, including four subsets, length covering 16k to 128k, 695 samples in total, the annotations are human-driven. We evaluate commercial and open-source models on CNNSum and conduct a detailed analysis. Based on the observations, we further conduct fine-tuning exploration with short-context summary data. In our study: (1) GPT-4o underperformed, due to excessive subjective commentary. (2) Currently, long-context summarization mainly relies on memory ability, small LLMs with stable longer context lengths are the most cost-effective. Using long data concatenated from short-context summaries makes a significant improvement. (3) Prompt templates may cause a large performance gap but can be mitigated through fine-tuning. (4) Fine-tuned Chat or Instruction versions may harm the Base model and further fine-tuning cannot bridge performance gap. (5) while models with RoPE base scaling exhibit strong extrapolation potential, their performance may vary significantly when combined with other interpolation methods and need careful selection. (6) CNNSum provides more reliable and insightful evaluation results than other benchmarks. We release CNNSum to advance research in this field (https://github.com/CxsGhost/CNNSum).
♻ ☆ Language Ranker: A Metric for Quantifying LLM Performance Across High and Low-Resource Languages AAAI 2025
The development of Large Language Models (LLMs) relies on extensive text corpora, which are often unevenly distributed across languages. This imbalance results in LLMs performing significantly better on high-resource languages like English, German, and French, while their capabilities in low-resource languages remain inadequate. Currently, there is a lack of quantitative methods to evaluate the performance of LLMs in these low-resource languages. To address this gap, we propose the Language Ranker, an intrinsic metric designed to benchmark and rank languages based on LLM performance using internal representations. By comparing the LLM's internal representation of various languages against a baseline derived from English, we can assess the model's multilingual capabilities in a robust and language-agnostic manner. Our analysis reveals that high-resource languages exhibit higher similarity scores with English, demonstrating superior performance, while low-resource languages show lower similarity scores, underscoring the effectiveness of our metric in assessing language-specific capabilities. Besides, the experiments show that there is a strong correlation between the LLM's performance in different languages and the proportion of those languages in its pre-training corpus. These insights underscore the efficacy of the Language Ranker as a tool for evaluating LLM performance across different languages, particularly those with limited resources.
comment: Accepted by AAAI 2025 (Social Impact Track)
♻ ☆ Efficient Prompt Tuning by Multi-Space Projection and Prompt Fusion
Prompt tuning is a promising method to fine-tune a pre-trained language model without retraining its large-scale parameters. Instead, it attaches a soft prompt to the input text, whereby downstream tasks can be well adapted by merely learning the embeddings of prompt tokens. Nevertheless, existing methods still suffer from two challenges: (i) they are hard to balance accuracy and efficiency. A longer (shorter) soft prompt generally leads to a better(worse) accuracy but at the cost of more (less) training time. (ii)The performance may not be consistent when adapting to different downstream tasks. We attribute it to the same embedding space but responsible for different requirements of downstream tasks. To address these issues, we propose an Efficient Prompt Tuning method (EPT) by multi-space projection and prompt fusion. Specifically, it decomposes a given soft prompt into a shorter prompt and two low-rank matrices, significantly reducing the training time. Accuracy is also enhanced by leveraging low-rank matrices and the short prompt as additional knowledge sources to enrich the semantics of the original short prompt. In addition, we project the soft prompt into multiple subspaces to improve the performance consistency, and then adaptively learn the combination weights of different spaces through a gating network. Experiments on 13 natural language processing downstream tasks show that our method significantly and consistently outperforms 11 comparison methods with the relative percentage of improvements up to 12.9%, and training time decreased by 14%.
♻ ☆ GameArena: Evaluating LLM Reasoning through Live Computer Games
Evaluating the reasoning abilities of large language models (LLMs) is challenging. Existing benchmarks often depend on static datasets, which are vulnerable to data contamination and may get saturated over time, or on binary live human feedback that conflates reasoning with other abilities. As the most prominent dynamic benchmark, Chatbot Arena evaluates open-ended questions in real-world settings, but lacks the granularity in assessing specific reasoning capabilities. We introduce GameArena, a dynamic benchmark designed to evaluate LLM reasoning capabilities through interactive gameplay with humans. GameArena consists of three games designed to test specific reasoning capabilities (e.g., deductive and inductive reasoning), while keeping participants entertained and engaged. We analyze the gaming data retrospectively to uncover the underlying reasoning processes of LLMs and measure their fine-grained reasoning capabilities. We collect over 2000 game sessions and provide detailed assessments of various reasoning capabilities for five state-of-the-art LLMs. Our user study with 100 participants suggests that GameArena improves user engagement compared to Chatbot Arena. For the first time, GameArena enables the collection of step-by-step LLM reasoning data in the wild.
♻ ☆ ShifCon: Enhancing Non-Dominant Language Capabilities with a Shift-based Contrastive Framework
Although fine-tuning Large Language Models (LLMs) with multilingual data can rapidly enhance the multilingual capabilities of LLMs, they still exhibit a performance gap between the dominant language (e.g., English) and non-dominant ones due to the imbalance of training data across languages. To further enhance the performance of non-dominant languages, we propose ShifCon, a Shift-based Contrastive framework that aligns the internal forward process of other languages toward that of the dominant one. Specifically, it shifts the representations of non-dominant languages into the dominant language subspace, allowing them to access relatively rich information encoded in the model parameters. The enriched representations are then shifted back into their original language subspace before generation. Moreover, we introduce a subspace distance metric to pinpoint the optimal layer area for shifting representations and employ multilingual contrastive learning to further enhance the alignment of representations within this area. Experiments demonstrate that our ShifCon framework significantly enhances the performance of non-dominant languages, particularly for low-resource ones. Further analysis offers extra insights to verify the effectiveness of ShifCon and propel future research
comment: 23 pages, 11 figures
♻ ☆ NushuRescue: Revitalization of the Endangered Nushu Language with AI COLING 2025
The preservation and revitalization of endangered and extinct languages is a meaningful endeavor, conserving cultural heritage while enriching fields like linguistics and anthropology. However, these languages are typically low-resource, making their reconstruction labor-intensive and costly. This challenge is exemplified by Nushu, a rare script historically used by Yao women in China for self-expression within a patriarchal society. To address this challenge, we introduce NushuRescue, an AI-driven framework designed to train large language models (LLMs) on endangered languages with minimal data. NushuRescue automates evaluation and expands target corpora to accelerate linguistic revitalization. As a foundational component, we developed NCGold, a 500-sentence Nushu-Chinese parallel corpus, the first publicly available dataset of its kind. Leveraging GPT-4-Turbo, with no prior exposure to Nushu and only 35 short examples from NCGold, NushuRescue achieved 48.69% translation accuracy on 50 withheld sentences and generated NCSilver, a set of 98 newly translated modern Chinese sentences of varying lengths. A sample of both NCGold and NCSilver is included in the Supplementary Materials. Additionally, we developed FastText-based and Seq2Seq models to further support research on Nushu. NushuRescue provides a versatile and scalable tool for the revitalization of endangered languages, minimizing the need for extensive human input.
comment: Accepted to COLING 2025
♻ ☆ Unveiling Language Competence Neurons: A Psycholinguistic Approach to Model Interpretability
As large language models (LLMs) advance in their linguistic capacity, understanding how they capture aspects of language competence remains a significant challenge. This study therefore employs psycholinguistic paradigms in English, which are well-suited for probing deeper cognitive aspects of language processing, to explore neuron-level representations in language model across three tasks: sound-shape association, sound-gender association, and implicit causality. Our findings indicate that while GPT-2-XL struggles with the sound-shape task, it demonstrates human-like abilities in both sound-gender association and implicit causality. Targeted neuron ablation and activation manipulation reveal a crucial relationship: When GPT-2-XL displays a linguistic ability, specific neurons correspond to that competence; conversely, the absence of such an ability indicates a lack of specialized neurons. This study is the first to utilize psycholinguistic experiments to investigate deep language competence at the neuron level, providing a new level of granularity in model interpretability and insights into the internal mechanisms driving language ability in the transformer-based LLM.
♻ ☆ LABOR-LLM: Language-Based Occupational Representations with Large Language Models
Vafa et al. (2024) introduced a transformer-based econometric model, CAREER, that predicts a worker's next job as a function of career history (an "occupation model"). CAREER was initially estimated ("pre-trained") using a large, unrepresentative resume dataset, which served as a "foundation model," and parameter estimation was continued ("fine-tuned") using data from a representative survey. CAREER had better predictive performance than benchmarks. This paper considers an alternative where the resume-based foundation model is replaced by a large language model (LLM). We convert tabular data from the survey into text files that resemble resumes and fine-tune the LLMs using these text files with the objective to predict the next token (word). The resulting fine-tuned LLM is used as an input to an occupation model. Its predictive performance surpasses all prior models. We demonstrate the value of fine-tuning and further show that by adding more career data from a different population, fine-tuning smaller LLMs surpasses the performance of fine-tuning larger models.
♻ ☆ Dynamic Universal Approximation Theory: The Basic Theory for Transformer-based Large Language Models
Language models have emerged as a critical area of focus in artificial intelligence, particularly with the introduction of groundbreaking innovations like ChatGPT. Large-scale Transformer networks have quickly become the leading approach for advancing natural language processing algorithms. Built on the Transformer architecture, these models enable interactions that closely mimic human communication and, equipped with extensive knowledge, can even assist in guiding human tasks. Despite their impressive capabilities and growing complexity, a key question remains-the theoretical foundations of large language models (LLMs). What makes Transformer so effective for powering intelligent language applications, such as translation and coding? What underlies LLMs' ability for In-Context Learning (ICL)? How does the LoRA scheme enhance the fine-tuning of LLMs? And what supports the practicality of pruning LLMs? To address these critical questions and explore the technological strategies within LLMs, we leverage the Universal Approximation Theory (UAT) to offer a theoretical backdrop, shedding light on the mechanisms that underpin these advancements.
♻ ☆ Dynamic Demonstrations Controller for In-Context Learning
In-context learning (ICL) is a new paradigm for natural language processing (NLP), where a large language model (LLM) observes a small number of demonstrations and a test instance as its input, and directly makes predictions without updating model parameters. Previous studies have revealed that ICL is sensitive to the selection and the ordering of demonstrations. However, there are few studies regarding the impact of the demonstration number on the ICL performance within a limited input length of LLM, because it is commonly believed that the number of demonstrations is positively correlated with model performance. In this paper, we found this conclusion does not always hold true. Through pilot experiments, we discover that increasing the number of demonstrations does not necessarily lead to improved performance. Building upon this insight, we propose a Dynamic Demonstrations Controller (D$^2$Controller), which can improve the ICL performance by adjusting the number of demonstrations dynamically. The experimental results show that D$^2$Controller yields a 4.6% relative improvement on ten different sizes of LLMs across ten datasets. Moreover, we also extend our method to previous ICL models and achieve competitive results.
♻ ☆ Training Large Language Models to Reason in a Continuous Latent Space
Large language models (LLMs) are restricted to reason in the "language space", where they typically express the reasoning process with a chain-of-thought (CoT) to solve a complex reasoning problem. However, we argue that language space may not always be optimal for reasoning. For example, most word tokens are primarily for textual coherence and not essential for reasoning, while some critical tokens require complex planning and pose huge challenges to LLMs. To explore the potential of LLM reasoning in an unrestricted latent space instead of using natural language, we introduce a new paradigm Coconut (Chain of Continuous Thought). We utilize the last hidden state of the LLM as a representation of the reasoning state (termed "continuous thought"). Rather than decoding this into a word token, we feed it back to the LLM as the subsequent input embedding directly in the continuous space. Experiments show that Coconut can effectively augment the LLM on several reasoning tasks. This novel latent reasoning paradigm leads to emergent advanced reasoning patterns: the continuous thought can encode multiple alternative next reasoning steps, allowing the model to perform a breadth-first search (BFS) to solve the problem, rather than prematurely committing to a single deterministic path like CoT. Coconut outperforms CoT in certain logical reasoning tasks that require substantial backtracking during planning, with fewer thinking tokens during inference. These findings demonstrate the promise of latent reasoning and offer valuable insights for future research.
♻ ☆ Large Language Models and Cognitive Science: A Comprehensive Review of Similarities, Differences, and Challenges
This comprehensive review explores the intersection of Large Language Models (LLMs) and cognitive science, examining similarities and differences between LLMs and human cognitive processes. We analyze methods for evaluating LLMs cognitive abilities and discuss their potential as cognitive models. The review covers applications of LLMs in various cognitive fields, highlighting insights gained for cognitive science research. We assess cognitive biases and limitations of LLMs, along with proposed methods for improving their performance. The integration of LLMs with cognitive architectures is examined, revealing promising avenues for enhancing artificial intelligence (AI) capabilities. Key challenges and future research directions are identified, emphasizing the need for continued refinement of LLMs to better align with human cognition. This review provides a balanced perspective on the current state and future potential of LLMs in advancing our understanding of both artificial and human intelligence.
comment: 10 pages, 1 figure
♻ ☆ Plentiful Jailbreaks with String Compositions NeurIPS
Large language models (LLMs) remain vulnerable to a slew of adversarial attacks and jailbreaking methods. One common approach employed by white-hat attackers, or red-teamers, is to process model inputs and outputs using string-level obfuscations, which can include leetspeak, rotary ciphers, Base64, ASCII, and more. Our work extends these encoding-based attacks by unifying them in a framework of invertible string transformations. With invertibility, we can devise arbitrary string compositions, defined as sequences of transformations, that we can encode and decode end-to-end programmatically. We devise a automated best-of-n attack that samples from a combinatorially large number of string compositions. Our jailbreaks obtain competitive attack success rates on several leading frontier models when evaluated on HarmBench, highlighting that encoding-based attacks remain a persistent vulnerability even in advanced LLMs.
comment: NeurIPS SoLaR Workshop 2024
♻ ☆ Woodpecker: Hallucination Correction for Multimodal Large Language Models SC
Hallucination is a big shadow hanging over the rapidly evolving Multimodal Large Language Models (MLLMs), referring to the phenomenon that the generated text is inconsistent with the image content. In order to mitigate hallucinations, existing studies mainly resort to an instruction-tuning manner that requires retraining the models with specific data. In this paper, we pave a different way, introducing a training-free method named Woodpecker. Like a woodpecker heals trees, it picks out and corrects hallucinations from the generated text. Concretely, Woodpecker consists of five stages: key concept extraction, question formulation, visual knowledge validation, visual claim generation, and hallucination correction. Implemented in a post-remedy manner, Woodpecker can easily serve different MLLMs, while being interpretable by accessing intermediate outputs of the five stages. We evaluate Woodpecker both quantitatively and qualitatively and show the huge potential of this new paradigm. On the POPE benchmark, our method obtains a 30.66%/24.33% improvement in accuracy over the baseline MiniGPT-4/mPLUG-Owl. The source code is released at https://github.com/BradyFU/Woodpecker.
comment: Accepted by Science China Information Sciences (SCIS)
♻ ☆ Enhancing Relation Extraction via Supervised Rationale Verification and Feedback AAAI 2025
Despite the rapid progress that existing automated feedback methods have made in correcting the output of large language models (LLMs), these methods cannot be well applied to the relation extraction (RE) task due to their designated feedback objectives and correction manner. To address this problem, we propose a novel automated feedback framework for RE, which presents a rationale supervisor to verify the rationale and provides re-selected demonstrations as feedback to correct the initial prediction. Specifically, we first design a causal intervention and observation method to collect biased/unbiased rationales for contrastive training the rationale supervisor. Then, we present a verification-feedback-correction procedure to iteratively enhance LLMs' capability of handling the RE task. Extensive experiments prove that our proposed framework significantly outperforms existing methods.
comment: Accepted to AAAI 2025, camera ready version
♻ ☆ Fundamental Limits of Prompt Compression: A Rate-Distortion Framework for Black-Box Language Models NeurIPS 2024
We formalize the problem of prompt compression for large language models (LLMs) and present a framework to unify token-level prompt compression methods which create hard prompts for black-box models. We derive the distortion-rate function for this setup as a linear program, and provide an efficient algorithm to compute this fundamental limit via the dual of the linear program. Using the distortion-rate function as the baseline, we study the performance of existing compression schemes on a synthetic dataset consisting of prompts generated from a Markov chain, natural language queries, and their respective answers. Our empirical analysis demonstrates the criticality of query-aware prompt compression, where the compressor has knowledge of the downstream task/query for the black-box LLM. We show that there is a large gap between the performance of current prompt compression methods and the optimal strategy, and propose Adaptive QuerySelect, a query-aware, variable-rate adaptation of a prior work to close the gap. We extend our experiments to a small natural language dataset to further confirm our findings on our synthetic dataset.
comment: 42 pages, 17 figures. Accepted to NeurIPS 2024
♻ ☆ Technical Report: Enhancing LLM Reasoning with Reward-guided Tree Search
Recently, test-time scaling has garnered significant attention from the research community, largely due to the substantial advancements of the o1 model released by OpenAI. By allocating more computational resources during the inference phase, large language models~(LLMs) can extensively explore the solution space by generating more thought tokens or diverse solutions, thereby producing more accurate responses. However, developing an o1-like reasoning approach is challenging, and researchers have been making various attempts to advance this open area of research. In this paper, we present a preliminary exploration into enhancing the reasoning abilities of LLMs through reward-guided tree search algorithms. This framework is implemented by integrating the policy model, reward model, and search algorithm. It is primarily constructed around a tree search algorithm, where the policy model navigates a dynamically expanding tree guided by a specially trained reward model. We thoroughly explore various design considerations necessary for implementing this framework and provide a detailed report of the technical aspects. To assess the effectiveness of our approach, we focus on mathematical reasoning tasks and conduct extensive evaluations on four challenging datasets, significantly enhancing the reasoning abilities of LLMs.
comment: Technical Report on Slow Thinking with LLMs: I
♻ ☆ ENTP: Encoder-only Next Token Prediction
Next-token prediction is conventionally done using decoder-only Transformers with causal attention, as this approach allows for efficient reuse of keys and values. What if we were not compute-limited, should we still use decoder-only Transformers? In this work, we introduce Encoder-only Next Token Prediction (ENTP). We use small scale experiments to explore the differences between ENTP and decoders, highlighting potential advantages of ENTP in setting with unbounded compute. We introduce the Count3 task and show, both theoretically and experimentally, that while ENTP can perform this task easily, a decoder-only Transformer cannot. Finally, we empirically demonstrate ENTP's superior performance across various synthetic tasks, such as length generalization and in-context learning.
Information Retrieval 18
☆ Preference Discerning with LLM-Enhanced Generative Retrieval
Sequential recommendation systems aim to provide personalized recommendations for users based on their interaction history. To achieve this, they often incorporate auxiliary information, such as textual descriptions of items and auxiliary tasks, like predicting user preferences and intent. Despite numerous efforts to enhance these models, they still suffer from limited personalization. To address this issue, we propose a new paradigm, which we term preference discerning. In preference dscerning, we explicitly condition a generative sequential recommendation system on user preferences within its context. To this end, we generate user preferences using Large Language Models (LLMs) based on user reviews and item-specific data. To evaluate preference discerning capabilities of sequential recommendation systems, we introduce a novel benchmark that provides a holistic evaluation across various scenarios, including preference steering and sentiment following. We assess current state-of-the-art methods using our benchmark and show that they struggle to accurately discern user preferences. Therefore, we propose a new method named Mender ($\textbf{M}$ultimodal Prefer$\textbf{en}$ce $\textbf{d}$iscern$\textbf{er}$), which improves upon existing methods and achieves state-of-the-art performance on our benchmark. Our results show that Mender can be effectively guided by human preferences even though they have not been observed during training, paving the way toward more personalized sequential recommendation systems. We will open-source the code and benchmarks upon publication.
comment: 11 pages + references and appendix
☆ Leveraging Graph-RAG and Prompt Engineering to Enhance LLM-Based Automated Requirement Traceability and Compliance Checks
Ensuring that Software Requirements Specifications (SRS) align with higher-level organizational or national requirements is vital, particularly in regulated environments such as finance and aerospace. In these domains, maintaining consistency, adhering to regulatory frameworks, minimizing errors, and meeting critical expectations are essential for the reliable functioning of systems. The widespread adoption of large language models (LLMs) highlights their immense potential, yet there remains considerable scope for improvement in retrieving relevant information and enhancing reasoning capabilities. This study demonstrates that integrating a robust Graph-RAG framework with advanced prompt engineering techniques, such as Chain of Thought and Tree of Thought, can significantly enhance performance. Compared to baseline RAG methods and simple prompting strategies, this approach delivers more accurate and context-aware results. While this method demonstrates significant improvements in performance, it comes with challenges. It is both costly and more complex to implement across diverse contexts, requiring careful adaptation to specific scenarios. Additionally, its effectiveness heavily relies on having complete and accurate input data, which may not always be readily available, posing further limitations to its scalability and practicality.
☆ AltFS: Agency-light Feature Selection with Large Language Models in Deep Recommender Systems
Feature selection is crucial in recommender systems for improving model efficiency and predictive performance. Traditional methods rely on agency models, such as decision trees or neural networks, to estimate feature importance. However, this approach is inherently limited, as the agency models may fail to learn effectively in all scenarios due to suboptimal training conditions (e.g., feature collinearity, high-dimensional sparsity, and data insufficiency). In this paper, we propose AltFS, an Agency-light Feature Selection method for deep recommender systems. AltFS integrates semantic reasoning from Large Language Models (LLMs) with task-specific learning from agency models. Initially, LLMs will generate a semantic ranking of feature importance, which is then refined by an agency model, combining world knowledge with task-specific insights. Extensive experiments on three public datasets from real-world recommender platforms demonstrate the effectiveness of AltFS. Our code is publicly available for reproducibility.
comment: under review
☆ InvDiff: Invariant Guidance for Bias Mitigation in Diffusion Models KDD 2025
As one of the most successful generative models, diffusion models have demonstrated remarkable efficacy in synthesizing high-quality images. These models learn the underlying high-dimensional data distribution in an unsupervised manner. Despite their success, diffusion models are highly data-driven and prone to inheriting the imbalances and biases present in real-world data. Some studies have attempted to address these issues by designing text prompts for known biases or using bias labels to construct unbiased data. While these methods have shown improved results, real-world scenarios often contain various unknown biases, and obtaining bias labels is particularly challenging. In this paper, we emphasize the necessity of mitigating bias in pre-trained diffusion models without relying on auxiliary bias annotations. To tackle this problem, we propose a framework, InvDiff, which aims to learn invariant semantic information for diffusion guidance. Specifically, we propose identifying underlying biases in the training data and designing a novel debiasing training objective. Then, we employ a lightweight trainable module that automatically preserves invariant semantic information and uses it to guide the diffusion model's sampling process toward unbiased outcomes simultaneously. Notably, we only need to learn a small number of parameters in the lightweight learnable module without altering the pre-trained diffusion model. Furthermore, we provide a theoretical guarantee that the implementation of InvDiff is equivalent to reducing the error upper bound of generalization. Extensive experimental results on three publicly available benchmarks demonstrate that InvDiff effectively reduces biases while maintaining the quality of image generation. Our code is available at https://github.com/Hundredl/InvDiff.
comment: KDD 2025
☆ NyayaAnumana & INLegalLlama: The Largest Indian Legal Judgment Prediction Dataset and Specialized Language Model for Enhanced Decision Analysis COLING 2025
The integration of artificial intelligence (AI) in legal judgment prediction (LJP) has the potential to transform the legal landscape, particularly in jurisdictions like India, where a significant backlog of cases burdens the legal system. This paper introduces NyayaAnumana, the largest and most diverse corpus of Indian legal cases compiled for LJP, encompassing a total of 7,02,945 preprocessed cases. NyayaAnumana, which combines the words "Nyay" (judgment) and "Anuman" (prediction or inference) respectively for most major Indian languages, includes a wide range of cases from the Supreme Court, High Courts, Tribunal Courts, District Courts, and Daily Orders and, thus, provides unparalleled diversity and coverage. Our dataset surpasses existing datasets like PredEx and ILDC, offering a comprehensive foundation for advanced AI research in the legal domain. In addition to the dataset, we present INLegalLlama, a domain-specific generative large language model (LLM) tailored to the intricacies of the Indian legal system. It is developed through a two-phase training approach over a base LLaMa model. First, Indian legal documents are injected using continual pretraining. Second, task-specific supervised finetuning is done. This method allows the model to achieve a deeper understanding of legal contexts. Our experiments demonstrate that incorporating diverse court data significantly boosts model accuracy, achieving approximately 90% F1-score in prediction tasks. INLegalLlama not only improves prediction accuracy but also offers comprehensible explanations, addressing the need for explainability in AI-assisted legal decisions.
comment: Accepted on COLING 2025
☆ Augmenting Sequential Recommendation with Balanced Relevance and Diversity AAAI 2025
By generating new yet effective data, data augmentation has become a promising method to mitigate the data sparsity problem in sequential recommendation. Existing works focus on augmenting the original data but rarely explore the issue of imbalanced relevance and diversity for augmented data, leading to semantic drift problems or limited performance improvements. In this paper, we propose a novel Balanced data Augmentation Plugin for Sequential Recommendation (BASRec) to generate data that balance relevance and diversity. BASRec consists of two modules: Single-sequence Augmentation and Cross-sequence Augmentation. The former leverages the randomness of the heuristic operators to generate diverse sequences for a single user, after which the diverse and the original sequences are fused at the representation level to obtain relevance. Further, we devise a reweighting strategy to enable the model to learn the preferences based on the two properties adaptively. The Cross-sequence Augmentation performs nonlinear mixing between different sequence representations from two directions. It produces virtual sequence representations that are diverse enough but retain the vital semantics of the original sequences. These two modules enhance the model to discover fine-grained preferences knowledge from single-user and cross-user perspectives. Extensive experiments verify the effectiveness of BASRec. The average improvement is up to 72.0% on GRU4Rec, 33.8% on SASRec, and 68.5% on FMLP-Rec. We demonstrate that BASRec generates data with a better balance between relevance and diversity than existing methods. The source code is available at https://github.com/KingGugu/BASRec.
comment: Accepted by AAAI 2025
☆ Large Language Models for Scholarly Ontology Generation: An Extensive Analysis in the Engineering Field
Ontologies of research topics are crucial for structuring scientific knowledge, enabling scientists to navigate vast amounts of research, and forming the backbone of intelligent systems such as search engines and recommendation systems. However, manual creation of these ontologies is expensive, slow, and often results in outdated and overly general representations. As a solution, researchers have been investigating ways to automate or semi-automate the process of generating these ontologies. This paper offers a comprehensive analysis of the ability of large language models (LLMs) to identify semantic relationships between different research topics, which is a critical step in the development of such ontologies. To this end, we developed a gold standard based on the IEEE Thesaurus to evaluate the task of identifying four types of relationships between pairs of topics: broader, narrower, same-as, and other. Our study evaluates the performance of seventeen LLMs, which differ in scale, accessibility (open vs. proprietary), and model type (full vs. quantised), while also assessing four zero-shot reasoning strategies. Several models have achieved outstanding results, including Mixtral-8x7B, Dolphin-Mistral-7B, and Claude 3 Sonnet, with F1-scores of 0.847, 0.920, and 0.967, respectively. Furthermore, our findings demonstrate that smaller, quantised models, when optimised through prompt engineering, can deliver performance comparable to much larger proprietary models, while requiring significantly fewer computational resources.
comment: submitted to Information Processing & Management
☆ Exploring Multidimensional Checkworthiness: Designing AI-assisted Claim Prioritization for Human Fact-checkers
Given the massive volume of potentially false claims circulating online, claim prioritization is essential in allocating limited human resources available for fact-checking. In this study, we perceive claim prioritization as an information retrieval (IR) task: just as multidimensional IR relevance, with many factors influencing which search results a user deems relevant, checkworthiness is also multi-faceted, subjective, and even personal, with many factors influencing how fact-checkers triage and select which claims to check. Our study investigates both the multidimensional nature of checkworthiness and effective tool support to assist fact-checkers in claim prioritization. Methodologically, we pursue Research through Design combined with mixed-method evaluation. We develop an AI-assisted claim prioritization prototype as a probe to explore how fact-checkers use multidimensional checkworthiness factors in claim prioritization, simultaneously probing fact-checker needs while also exploring the design space to meet those needs. Our study with 16 professional fact-checkers investigates: 1) how participants assessed the relative importance of different checkworthy dimensions and apply different priorities in claim selection; 2) how they created customized GPT-based search filters and the corresponding benefits and limitations; and 3) their overall user experiences with our prototype. Our work makes a conceptual contribution between multidimensional IR relevance and fact-checking checkworthiness, with findings demonstrating the value of corresponding tooling support. Specifically, we uncovered a hierarchical prioritization strategy fact-checkers implicitly use, revealing an underexplored aspect of their workflow, with actionable design recommendations for improving claim triage across multi-dimensional checkworthiness and tailoring this process with LLM integration.
☆ Multimodal Difference Learning for Sequential Recommendation
Sequential recommendations have drawn significant attention in modeling the user's historical behaviors to predict the next item. With the booming development of multimodal data (e.g., image, text) on internet platforms, sequential recommendation also benefits from the incorporation of multimodal data. Most methods introduce modal features of items as side information and simply concatenates them to learn unified user interests. Nevertheless, these methods encounter the limitation in modeling multimodal differences. We argue that user interests and item relationships vary across different modalities. To address this problem, we propose a novel Multimodal Difference Learning framework for Sequential Recommendation, MDSRec for brevity. Specifically, we first explore the differences in item relationships by constructing modal-aware item relation graphs with behavior signal to enhance item representations. Then, to capture the differences in user interests across modalities, we design a interest-centralized attention mechanism to independently model user sequence representations in different modalities. Finally, we fuse the user embeddings from multiple modalities to achieve accurate item recommendation. Experimental results on five real-world datasets demonstrate the superiority of MDSRec over state-of-the-art baselines and the efficacy of multimodal difference learning.
☆ A Tutorial of Personalized Federated Recommender Systems: Recent Advances and Future Directions
Personalization stands as the cornerstone of recommender systems (RecSys), striving to sift out redundant information and offer tailor-made services for users. However, the conventional cloud-based RecSys necessitates centralized data collection, posing significant risks of user privacy breaches. In response to this challenge, federated recommender systems (FedRecSys) have emerged, garnering considerable attention. FedRecSys enable users to retain personal data locally and solely share model parameters with low privacy sensitivity for global model training, significantly bolstering the system's privacy protection capabilities. Within the distributed learning framework, the pronounced non-iid nature of user behavior data introduces fresh hurdles to federated optimization. Meanwhile, the ability of federated learning to concurrently learn multiple models presents an opportunity for personalized user modeling. Consequently, the development of personalized FedRecSys (PFedRecSys) is crucial and holds substantial significance. This tutorial seeks to provide an introduction to PFedRecSys, encompassing (1) an overview of existing studies on PFedRecSys, (2) a comprehensive taxonomy of PFedRecSys spanning four pivotal research directions-client-side adaptation, server-side aggregation, communication efficiency, privacy and protection, and (3) exploration of open challenges and promising future directions in PFedRecSys. This tutorial aims to establish a robust foundation and spark new perspectives for subsequent exploration and practical implementations in the evolving realm of RecSys.
comment: A technical tutorial will appear at The Web Conference 2025
☆ Cluster-Enhanced Federated Graph Neural Network for Recommendation
Personal interaction data can be effectively modeled as individual graphs for each user in recommender systems.Graph Neural Networks (GNNs)-based recommendation techniques have become extremely popular since they can capture high-order collaborative signals between users and items by aggregating the individual graph into a global interactive graph.However, this centralized approach inherently poses a threat to user privacy and security. Recently, federated GNN-based recommendation techniques have emerged as a promising solution to mitigate privacy concerns. Nevertheless, current implementations either limit on-device training to an unaccompanied individual graphs or necessitate reliance on an extra third-party server to touch other individual graphs, which also increases the risk of privacy leakage. To address this challenge, we propose a Cluster-enhanced Federated Graph Neural Network framework for Recommendation, named CFedGR, which introduces high-order collaborative signals to augment individual graphs in a privacy preserving manner. Specifically, the server clusters the pretrained user representations to identify high-order collaborative signals. In addition, two efficient strategies are devised to reduce communication between devices and the server. Extensive experiments on three benchmark datasets validate the effectiveness of our proposed methods.
☆ RALI@TREC iKAT 2024: Achieving Personalization via Retrieval Fusion in Conversational Search
The Recherche Appliquee en Linguistique Informatique (RALI) team participated in the 2024 TREC Interactive Knowledge Assistance (iKAT) Track. In personalized conversational search, effectively capturing a user's complex search intent requires incorporating both contextual information and key elements from the user profile into query reformulation. The user profile often contains many relevant pieces, and each could potentially complement the user's information needs. It is difficult to disregard any of them, whereas introducing an excessive number of these pieces risks drifting from the original query and hinders search performance. This is a challenge we denote as over-personalization. To address this, we propose different strategies by fusing ranking lists generated from the queries with different levels of personalization.
comment: Work presented at NIST Text Retrieval Conference 2024. https://www.nist.gov/news-events/events/2024/11/trec2024
☆ jina-clip-v2: Multilingual Multimodal Embeddings for Text and Images
Contrastive Language-Image Pretraining (CLIP) is a highly effective method for aligning images and texts in a shared embedding space. These models are widely used for tasks such as cross-modal information retrieval and multi-modal understanding. However, CLIP models often struggle with text-only tasks, underperforming compared to specialized text models. This performance disparity forces retrieval systems to rely on separate models for text-only and multi-modal tasks. In this work, we build upon our previous model, jina-clip-v1, by introducing a refined framework that utilizes multi-task, multi-stage contrastive learning across multiple languages, coupled with an improved training recipe to enhance text-only retrieval. The resulting model, jina-clip-v2, outperforms its predecessor on text-only and multimodal tasks, while adding multilingual support, better understanding of complex visual documents and efficiency gains thanks to Matryoshka Representation Learning and vector truncation. The model performs comparably to the state-of-the-art in both multilingual-multimodal and multilingual text retrieval benchmarks, addressing the challenge of unifying text-only and multi-modal retrieval systems.
comment: 21 pages, 1-10 main paper, 10-12 refs, 12-21 benchmarks
☆ Reducing Popularity Influence by Addressing Position Bias
Position bias poses a persistent challenge in recommender systems, with much of the existing research focusing on refining ranking relevance and driving user engagement. However, in practical applications, the mitigation of position bias does not always result in detectable short-term improvements in ranking relevance. This paper provides an alternative, practically useful view of what position bias reduction methods can achieve. It demonstrates that position debiasing can spread visibility and interactions more evenly across the assortment, effectively reducing a skew in the popularity of items induced by the position bias through a feedback loop. We offer an explanation of how position bias affects item popularity. This includes an illustrative model of the item popularity histogram and the effect of the position bias on its skewness. Through offline and online experiments on our large-scale e-commerce platform, we show that position debiasing can significantly improve assortment utilization, without any degradation in user engagement or financial metrics. This makes the ranking fairer and helps attract more partners or content providers, benefiting the customers and the business in the long term.
♻ ☆ CURE: A dataset for Clinical Understanding & Retrieval Evaluation
Given the dominance of dense retrievers that do not generalize well beyond their training dataset distributions, domain-specific test sets are essential in evaluating retrieval. There are few test datasets for retrieval systems intended for use by healthcare providers in a point-of-care setting. To fill this gap we have collaborated with medical professionals to create CURE, an ad-hoc retrieval test dataset for passage ranking with 2000 queries spanning 10 medical domains with a monolingual (English) and two cross-lingual (French/Spanish -> English) conditions. In this paper, we describe how CURE was constructed and provide baseline results to showcase its effectiveness as an evaluation tool. CURE is published with a Creative Commons Attribution Non Commercial 4.0 license and can be accessed on Hugging Face.
♻ ☆ Spatial-Temporal Federated Learning for Lifelong Person Re-identification on Distributed Edges
Data drift is a thorny challenge when deploying person re-identification (ReID) models into real-world devices, where the data distribution is significantly different from that of the training environment and keeps changing. To tackle this issue, we propose a federated spatial-temporal incremental learning approach, named FedSTIL, which leverages both lifelong learning and federated learning to continuously optimize models deployed on many distributed edge clients. Unlike previous efforts, FedSTIL aims to mine spatial-temporal correlations among the knowledge learnt from different edge clients. Specifically, the edge clients first periodically extract general representations of drifted data to optimize their local models. Then, the learnt knowledge from edge clients will be aggregated by centralized parameter server, where the knowledge will be selectively and attentively distilled from spatial- and temporal-dimension with carefully designed mechanisms. Finally, the distilled informative spatial-temporal knowledge will be sent back to correlated edge clients to further improve the recognition accuracy of each edge client with a lifelong learning method. Extensive experiments on a mixture of five real-world datasets demonstrate that our method outperforms others by nearly 4% in Rank-1 accuracy, while reducing communication cost by 62%. All implementation codes are publicly available on https://github.com/MSNLAB/Federated-Lifelong-Person-ReID
♻ ☆ Representation Learning with Large Language Models for Recommendation WWW'24
Recommender systems have seen significant advancements with the influence of deep learning and graph neural networks, particularly in capturing complex user-item relationships. However, these graph-based recommenders heavily depend on ID-based data, potentially disregarding valuable textual information associated with users and items, resulting in less informative learned representations. Moreover, the utilization of implicit feedback data introduces potential noise and bias, posing challenges for the effectiveness of user preference learning. While the integration of large language models (LLMs) into traditional ID-based recommenders has gained attention, challenges such as scalability issues, limitations in text-only reliance, and prompt input constraints need to be addressed for effective implementation in practical recommender systems. To address these challenges, we propose a model-agnostic framework RLMRec that aims to enhance existing recommenders with LLM-empowered representation learning. It proposes a recommendation paradigm that integrates representation learning with LLMs to capture intricate semantic aspects of user behaviors and preferences. RLMRec incorporates auxiliary textual signals, develops a user/item profiling paradigm empowered by LLMs, and aligns the semantic space of LLMs with the representation space of collaborative relational signals through a cross-view alignment framework. This work further establish a theoretical foundation demonstrating that incorporating textual signals through mutual information maximization enhances the quality of representations. In our evaluation, we integrate RLMRec with state-of-the-art recommender models, while also analyzing its efficiency and robustness to noise data. Our implementation codes are available at https://github.com/HKUDS/RLMRec.
comment: Published as a WWW'24 full paper
♻ ☆ Toward Reliable Ad-hoc Scientific Information Extraction: A Case Study on Two Materials Datasets
We explore the ability of GPT-4 to perform ad-hoc schema based information extraction from scientific literature. We assess specifically whether it can, with a basic prompting approach, replicate two existing material science datasets, given the manuscripts from which they were originally manually extracted. We employ materials scientists to perform a detailed manual error analysis to assess where the model struggles to faithfully extract the desired information, and draw on their insights to suggest research directions to address this broadly important task.
comment: Update on 12/11/2024: Added some relevant literature that we missed in previous version of the paper
Machine Learning 150
☆ GPD-1: Generative Pre-training for Driving
Modeling the evolutions of driving scenarios is important for the evaluation and decision-making of autonomous driving systems. Most existing methods focus on one aspect of scene evolution such as map generation, motion prediction, and trajectory planning. In this paper, we propose a unified Generative Pre-training for Driving (GPD-1) model to accomplish all these tasks altogether without additional fine-tuning. We represent each scene with ego, agent, and map tokens and formulate autonomous driving as a unified token generation problem. We adopt the autoregressive transformer architecture and use a scene-level attention mask to enable intra-scene bi-directional interactions. For the ego and agent tokens, we propose a hierarchical positional tokenizer to effectively encode both 2D positions and headings. For the map tokens, we train a map vector-quantized autoencoder to efficiently compress ego-centric semantic maps into discrete tokens. We pre-train our GPD-1 on the large-scale nuPlan dataset and conduct extensive experiments to evaluate its effectiveness. With different prompts, our GPD-1 successfully generalizes to various tasks without finetuning, including scene generation, traffic simulation, closed-loop simulation, map prediction, and motion planning. Code: https://github.com/wzzheng/GPD.
comment: Code is available at: https://github.com/wzzheng/GPD
☆ Generative Semantic Communication: Architectures, Technologies, and Applications
This paper delves into the applications of generative artificial intelligence (GAI) in semantic communication (SemCom) and presents a thorough study. Three popular SemCom systems enabled by classical GAI models are first introduced, including variational autoencoders, generative adversarial networks, and diffusion models. For each system, the fundamental concept of the GAI model, the corresponding SemCom architecture, and the associated literature review of recent efforts are elucidated. Then, a novel generative SemCom system is proposed by incorporating the cutting-edge GAI technology-large language models (LLMs). This system features two LLM-based AI agents at both the transmitter and receiver, serving as "brains" to enable powerful information understanding and content regeneration capabilities, respectively. This innovative design allows the receiver to directly generate the desired content, instead of recovering the bit stream, based on the coded semantic information conveyed by the transmitter. Therefore, it shifts the communication mindset from "information recovery" to "information regeneration" and thus ushers in a new era of generative SemCom. A case study on point-to-point video retrieval is presented to demonstrate the superiority of the proposed generative SemCom system, showcasing a 99.98% reduction in communication overhead and a 53% improvement in retrieval accuracy compared to the traditional communication system. Furthermore, four typical application scenarios for generative SemCom are delineated, followed by a discussion of three open issues warranting future investigation. In a nutshell, this paper provides a holistic set of guidelines for applying GAI in SemCom, paving the way for the efficient implementation of generative SemCom in future wireless networks.
comment: 18 pages, 8 figures
☆ DMin: Scalable Training Data Influence Estimation for Diffusion Models
Identifying the training data samples that most influence a generated image is a critical task in understanding diffusion models, yet existing influence estimation methods are constrained to small-scale or LoRA-tuned models due to computational limitations. As diffusion models scale up, these methods become impractical. To address this challenge, we propose DMin (Diffusion Model influence), a scalable framework for estimating the influence of each training data sample on a given generated image. By leveraging efficient gradient compression and retrieval techniques, DMin reduces storage requirements from 339.39 TB to only 726 MB and retrieves the top-k most influential training samples in under 1 second, all while maintaining performance. Our empirical results demonstrate DMin is both effective in identifying influential training samples and efficient in terms of computational and storage requirements.
comment: 14 pages, 6 figures, 8 tables. Under Review
☆ Multimodal Latent Language Modeling with Next-Token Diffusion
Multimodal generative models require a unified approach to handle both discrete data (e.g., text and code) and continuous data (e.g., image, audio, video). In this work, we propose Latent Language Modeling (LatentLM), which seamlessly integrates continuous and discrete data using causal Transformers. Specifically, we employ a variational autoencoder (VAE) to represent continuous data as latent vectors and introduce next-token diffusion for autoregressive generation of these vectors. Additionally, we develop $\sigma$-VAE to address the challenges of variance collapse, which is crucial for autoregressive modeling. Extensive experiments demonstrate the effectiveness of LatentLM across various modalities. In image generation, LatentLM surpasses Diffusion Transformers in both performance and scalability. When integrated into multimodal large language models, LatentLM provides a general-purpose interface that unifies multimodal generation and understanding. Experimental results show that LatentLM achieves favorable performance compared to Transfusion and vector quantized models in the setting of scaling up training tokens. In text-to-speech synthesis, LatentLM outperforms the state-of-the-art VALL-E 2 model in speaker similarity and robustness, while requiring 10x fewer decoding steps. The results establish LatentLM as a highly effective and scalable approach to advance large multimodal models.
☆ MNIST-Fraction: Enhancing Math Education with AI-Driven Fraction Detection and Analysis
Mathematics education, a crucial and basic field, significantly influences students' learning in related subjects and their future careers. Utilizing artificial intelligence to interpret and comprehend math problems in education is not yet fully explored. This is due to the scarcity of quality datasets and the intricacies of processing handwritten information. In this paper, we present a novel contribution to the field of mathematics education through the development of MNIST-Fraction, a dataset inspired by the renowned MNIST, specifically tailored for the recognition and understanding of handwritten math fractions. Our approach is the utilization of deep learning, specifically Convolutional Neural Networks (CNNs), for the recognition and understanding of handwritten math fractions to effectively detect and analyze fractions, along with their numerators and denominators. This capability is pivotal in calculating the value of fractions, a fundamental aspect of math learning. The MNIST-Fraction dataset is designed to closely mimic real-world scenarios, providing a reliable and relevant resource for AI-driven educational tools. Furthermore, we conduct a comprehensive comparison of our dataset with the original MNIST dataset using various classifiers, demonstrating the effectiveness and versatility of MNIST-Fraction in both detection and classification tasks. This comparative analysis not only validates the practical utility of our dataset but also offers insights into its potential applications in math education. To foster collaboration and further research within the computational and educational communities. Our work aims to bridge the gap in high-quality educational resources for math learning, offering a valuable tool for both educators and researchers in the field.
☆ FlowEdit: Inversion-Free Text-Based Editing Using Pre-Trained Flow Models
Editing real images using a pre-trained text-to-image (T2I) diffusion/flow model often involves inverting the image into its corresponding noise map. However, inversion by itself is typically insufficient for obtaining satisfactory results, and therefore many methods additionally intervene in the sampling process. Such methods achieve improved results but are not seamlessly transferable between model architectures. Here, we introduce FlowEdit, a text-based editing method for pre-trained T2I flow models, which is inversion-free, optimization-free and model agnostic. Our method constructs an ODE that directly maps between the source and target distributions (corresponding to the source and target text prompts) and achieves a lower transport cost than the inversion approach. This leads to state-of-the-art results, as we illustrate with Stable Diffusion 3 and FLUX. Code and examples are available on the project's webpage.
comment: Project's webpage at https://matankleiner.github.io/flowedit/
☆ Preference Discerning with LLM-Enhanced Generative Retrieval
Sequential recommendation systems aim to provide personalized recommendations for users based on their interaction history. To achieve this, they often incorporate auxiliary information, such as textual descriptions of items and auxiliary tasks, like predicting user preferences and intent. Despite numerous efforts to enhance these models, they still suffer from limited personalization. To address this issue, we propose a new paradigm, which we term preference discerning. In preference dscerning, we explicitly condition a generative sequential recommendation system on user preferences within its context. To this end, we generate user preferences using Large Language Models (LLMs) based on user reviews and item-specific data. To evaluate preference discerning capabilities of sequential recommendation systems, we introduce a novel benchmark that provides a holistic evaluation across various scenarios, including preference steering and sentiment following. We assess current state-of-the-art methods using our benchmark and show that they struggle to accurately discern user preferences. Therefore, we propose a new method named Mender ($\textbf{M}$ultimodal Prefer$\textbf{en}$ce $\textbf{d}$iscern$\textbf{er}$), which improves upon existing methods and achieves state-of-the-art performance on our benchmark. Our results show that Mender can be effectively guided by human preferences even though they have not been observed during training, paving the way toward more personalized sequential recommendation systems. We will open-source the code and benchmarks upon publication.
comment: 11 pages + references and appendix
☆ Numerical Analysis of HiPPO-LegS ODE for Deep State Space Models
In deep learning, the recently introduced state space models utilize HiPPO (High-order Polynomial Projection Operators) memory units to approximate continuous-time trajectories of input functions using ordinary differential equations (ODEs), and these techniques have shown empirical success in capturing long-range dependencies in long input sequences. However, the mathematical foundations of these ODEs, particularly the singular HiPPO-LegS (Legendre Scaled) ODE, and their corresponding numerical discretizations remain unexplored. In this work, we fill this gap by establishing that HiPPO-LegS ODE is well-posed despite its singularity, albeit without the freedom of arbitrary initial conditions, and by establishing convergence of the associated numerical discretization schemes for Riemann-integrable input functions.
☆ Adaptive Principal Components Allocation with the $\ell_{2,g}$-regularized Gaussian Graphical Model for Efficient Fine-Tuning Large Models
In this work, we propose a novel Parameter-Efficient Fine-Tuning (PEFT) approach based on Gaussian Graphical Models (GGMs), marking the first application of GGMs to PEFT tasks, to the best of our knowledge. The proposed method utilizes the $\ell_{2,g}$-norm to effectively select critical parameters and capture global dependencies. The resulting non-convex optimization problem is efficiently solved using a Block Coordinate Descent (BCD) algorithm. Experimental results on the GLUE benchmark [24] for fine-tuning RoBERTa-Base [18] demonstrate the effectiveness of the proposed approach, achieving competitive performance with significantly fewer trainable parameters. The code for this work is available at: https://github.com/jzheng20/Course projects.git.
☆ Preventing Conflicting Gradients in Neural Marked Temporal Point Processes
Neural Marked Temporal Point Processes (MTPP) are flexible models to capture complex temporal inter-dependencies between labeled events. These models inherently learn two predictive distributions: one for the arrival times of events and another for the types of events, also known as marks. In this study, we demonstrate that learning a MTPP model can be framed as a two-task learning problem, where both tasks share a common set of trainable parameters that are optimized jointly. We show that this often leads to the emergence of conflicting gradients during training, where task-specific gradients are pointing in opposite directions. When such conflicts arise, following the average gradient can be detrimental to the learning of each individual tasks, resulting in overall degraded performance. To overcome this issue, we introduce novel parametrizations for neural MTPP models that allow for separate modeling and training of each task, effectively avoiding the problem of conflicting gradients. Through experiments on multiple real-world event sequence datasets, we demonstrate the benefits of our framework compared to the original model formulations.
☆ SPACE-SUIT: An Artificial Intelligence based chromospheric feature extractor and classifier for SUIT
The Solar Ultraviolet Imaging Telescope(SUIT) onboard Aditya-L1 is an imager that observes the solar photosphere and chromosphere through observations in the wavelength range of 200-400 nm. A comprehensive understanding of the plasma and thermodynamic properties of chromospheric and photospheric morphological structures requires a large sample statistical study, necessitating the development of automatic feature detection methods. To this end, we develop the feature detection algorithm SPACE-SUIT: Solar Phenomena Analysis and Classification using Enhanced vision techniques for SUIT, to detect and classify the solar chromospheric features to be observed from SUIT's Mg II k filter. Specifically, we target plage regions, sunspots, filaments, and off-limb structures. SPACE uses You Only Look Once(YOLO), a neural network-based model to identify regions of interest. We train and validate SPACE using mock-SUIT images developed from Interface Region Imaging Spectrometer(IRIS) full-disk mosaic images in Mg II k line, while we also perform detection on Level-1 SUIT data. SPACE achieves an approximate precision of 0.788, recall 0.863 and MAP of 0.874 on the validation mock SUIT FITS dataset. Given the manual labeling of our dataset, we perform "self-validation" by applying statistical measures and Tamura features on the ground truth and predicted bounding boxes. We find the distributions of entropy, contrast, dissimilarity, and energy to show differences in the features. These differences are qualitatively captured by the detected regions predicted by SPACE and validated with the observed SUIT images, even in the absence of labeled ground truth. This work not only develops a chromospheric feature extractor but also demonstrates the effectiveness of statistical metrics and Tamura features for distinguishing chromospheric features, offering independent validation for future detection schemes.
☆ TURBOATTENTION: Efficient Attention Approximation For High Throughputs LLMs
Large language model (LLM) inference demands significant amount of computation and memory, especially in the key attention mechanism. While techniques, such as quantization and acceleration algorithms, like FlashAttention, have improved efficiency of the overall inference, they address different aspects of the problem: quantization focuses on weight-activation operations, while FlashAttention improves execution but requires high-precision formats. Recent Key-value (KV) cache quantization reduces memory bandwidth but still needs floating-point dequantization for attention operation. We present TurboAttention, a comprehensive approach to enable quantized execution of attention that simultaneously addresses both memory and computational efficiency. Our solution introduces two key innovations: FlashQ, a headwise attention quantization technique that enables both compression of KV cache and quantized execution of activation-activation multiplication, and Sparsity-based Softmax Approximation (SAS), which eliminates the need for dequantization to FP32 during exponentiation operation in attention. Experimental results demonstrate that TurboAttention achieves 1.2-1.8x speedup in attention, reduces the KV cache size by over 4.4x, and enables up to 2.37x maximum throughput over the FP16 baseline while outperforming state-of-the-art quantization and compression techniques across various datasets and models.
☆ GenPlan: Generative sequence models as adaptive planners AAAI 2025
Offline reinforcement learning has shown tremendous success in behavioral planning by learning from previously collected demonstrations. However, decision-making in multitask missions still presents significant challenges. For instance, a mission might require an agent to explore an unknown environment, discover goals, and navigate to them, even if it involves interacting with obstacles along the way. Such behavioral planning problems are difficult to solve due to: a) agents failing to adapt beyond the single task learned through their reward function, and b) the inability to generalize to new environments not covered in the training demonstrations, e.g., environments where all doors were unlocked in the demonstrations. Consequently, state-of-the-art decision making methods are limited to missions where the required tasks are well-represented in the training demonstrations and can be solved within a short (temporal) planning horizon. To address this, we propose GenPlan: a stochastic and adaptive planner that leverages discrete-flow models for generative sequence modeling, enabling sample-efficient exploration and exploitation. This framework relies on an iterative denoising procedure to generate a sequence of goals and actions. This approach captures multi-modal action distributions and facilitates goal and task discovery, thereby enhancing generalization to out-of-distribution tasks and environments, i.e., missions not part of the training data. We demonstrate the effectiveness of our method through multiple simulation environments. Notably, GenPlan outperforms the state-of-the-art methods by over 10% on adaptive planning tasks, where the agent adapts to multi-task missions while leveraging demonstrations on single-goal-reaching tasks.
comment: Accepted in AAAI 2025. Project page: https://aku02.github.io/projects/genplan/
☆ Underestimated Privacy Risks for Minority Populations in Large Language Model Unlearning
Large Language Models are trained on extensive datasets that often contain sensitive, human-generated information, raising significant concerns about privacy breaches. While certified unlearning approaches offer strong privacy guarantees, they rely on restrictive model assumptions that are not applicable to LLMs. As a result, various unlearning heuristics have been proposed, with the associated privacy risks assessed only empirically. The standard evaluation pipelines typically randomly select data for removal from the training set, apply unlearning techniques, and use membership inference attacks to compare the unlearned models against models retrained without the to-be-unlearned data. However, since every data point is subject to the right to be forgotten, unlearning should be considered in the worst-case scenario from the privacy perspective. Prior work shows that data outliers may exhibit higher memorization effects. Intuitively, they are harder to be unlearn and thus the privacy risk of unlearning them is underestimated in the current evaluation. In this paper, we leverage minority data to identify such a critical flaw in previously widely adopted evaluations. We substantiate this claim through carefully designed experiments, including unlearning canaries related to minority groups, inspired by privacy auditing literature. Using personally identifiable information as a representative minority identifier, we demonstrate that minority groups experience at least 20% more privacy leakage in most cases across six unlearning approaches, three MIAs, three benchmark datasets, and two LLMs of different scales. Given that the right to be forgotten should be upheld for every individual, we advocate for a more rigorous evaluation of LLM unlearning methods. Our minority-aware evaluation framework represents an initial step toward ensuring more equitable assessments of LLM unlearning efficacy.
☆ Grimm: A Plug-and-Play Perturbation Rectifier for Graph Neural Networks Defending against Poisoning Attacks
End-to-end training with global optimization have popularized graph neural networks (GNNs) for node classification, yet inadvertently introduced vulnerabilities to adversarial edge-perturbing attacks. Adversaries can exploit the inherent opened interfaces of GNNs' input and output, perturbing critical edges and thus manipulating the classification results. Current defenses, due to their persistent utilization of global-optimization-based end-to-end training schemes, inherently encapsulate the vulnerabilities of GNNs. This is specifically evidenced in their inability to defend against targeted secondary attacks. In this paper, we propose the Graph Agent Network (GAgN) to address the aforementioned vulnerabilities of GNNs. GAgN is a graph-structured agent network in which each node is designed as an 1-hop-view agent. Through the decentralized interactions between agents, they can learn to infer global perceptions to perform tasks including inferring embeddings, degrees and neighbor relationships for given nodes. This empowers nodes to filtering adversarial edges while carrying out classification tasks. Furthermore, agents' limited view prevents malicious messages from propagating globally in GAgN, thereby resisting global-optimization-based secondary attacks. We prove that single-hidden-layer multilayer perceptrons (MLPs) are theoretically sufficient to achieve these functionalities. Experimental results show that GAgN effectively implements all its intended capabilities and, compared to state-of-the-art defenses, achieves optimal classification accuracy on the perturbed datasets.
comment: 19 pages, 13 figures
☆ Watermarking Training Data of Music Generation Models
Generative Artificial Intelligence (Gen-AI) models are increasingly used to produce content across domains, including text, images, and audio. While these models represent a major technical breakthrough, they gain their generative capabilities from being trained on enormous amounts of human-generated content, which often includes copyrighted material. In this work, we investigate whether audio watermarking techniques can be used to detect an unauthorized usage of content to train a music generation model. We compare outputs generated by a model trained on watermarked data to a model trained on non-watermarked data. We study factors that impact the model's generation behaviour: the watermarking technique, the proportion of watermarked samples in the training set, and the robustness of the watermarking technique against the model's tokenizer. Our results show that audio watermarking techniques, including some that are imperceptible to humans, can lead to noticeable shifts in the model's outputs. We also study the robustness of a state-of-the-art watermarking technique to removal techniques.
☆ Training Data Reconstruction: Privacy due to Uncertainty?
Being able to reconstruct training data from the parameters of a neural network is a major privacy concern. Previous works have shown that reconstructing training data, under certain circumstances, is possible. In this work, we analyse such reconstructions empirically and propose a new formulation of the reconstruction as a solution to a bilevel optimisation problem. We demonstrate that our formulation as well as previous approaches highly depend on the initialisation of the training images $x$ to reconstruct. In particular, we show that a random initialisation of $x$ can lead to reconstructions that resemble valid training samples while not being part of the actual training dataset. Thus, our experiments on affine and one-hidden layer networks suggest that when reconstructing natural images, yet an adversary cannot identify whether reconstructed images have indeed been part of the set of training samples.
☆ MaestroMotif: Skill Design from Artificial Intelligence Feedback
Describing skills in natural language has the potential to provide an accessible way to inject human knowledge about decision-making into an AI system. We present MaestroMotif, a method for AI-assisted skill design, which yields high-performing and adaptable agents. MaestroMotif leverages the capabilities of Large Language Models (LLMs) to effectively create and reuse skills. It first uses an LLM's feedback to automatically design rewards corresponding to each skill, starting from their natural language description. Then, it employs an LLM's code generation abilities, together with reinforcement learning, for training the skills and combining them to implement complex behaviors specified in language. We evaluate MaestroMotif using a suite of complex tasks in the NetHack Learning Environment (NLE), demonstrating that it surpasses existing approaches in both performance and usability.
☆ Euclidean Fast Attention: Machine Learning Global Atomic Representations at Linear Cost
Long-range correlations are essential across numerous machine learning tasks, especially for data embedded in Euclidean space, where the relative positions and orientations of distant components are often critical for accurate predictions. Self-attention offers a compelling mechanism for capturing these global effects, but its quadratic complexity presents a significant practical limitation. This problem is particularly pronounced in computational chemistry, where the stringent efficiency requirements of machine learning force fields (MLFFs) often preclude accurately modeling long-range interactions. To address this, we introduce Euclidean fast attention (EFA), a linear-scaling attention-like mechanism designed for Euclidean data, which can be easily incorporated into existing model architectures. A core component of EFA are novel Euclidean rotary positional encodings (ERoPE), which enable efficient encoding of spatial information while respecting essential physical symmetries. We empirically demonstrate that EFA effectively captures diverse long-range effects, enabling EFA-equipped MLFFs to describe challenging chemical interactions for which conventional MLFFs yield incorrect results.
☆ Protecting Confidentiality, Privacy and Integrity in Collaborative Learning
A collaboration between dataset owners and model owners is needed to facilitate effective machine learning (ML) training. During this collaboration, however, dataset owners and model owners want to protect the confidentiality of their respective assets (i.e., datasets, models and training code), with the dataset owners also caring about the privacy of individual users whose data is in their datasets. Existing solutions either provide limited confidentiality for models and training code, or suffer from privacy issues due to collusion. We present Citadel++, a scalable collaborative ML training system designed to simultaneously protect the confidentiality of datasets, models and training code, as well as the privacy of individual users. Citadel++ enhances differential privacy techniques to safeguard the privacy of individual user data while maintaining model utility. By employing Virtual Machine-level Trusted Execution Environments (TEEs) and improved integrity protection techniques through various OS-level mechanisms, Citadel++ effectively preserves the confidentiality of datasets, models and training code, and enforces our privacy mechanisms even when the models and training code have been maliciously designed. Our experiments show that Citadel++ provides privacy, model utility and performance while adhering to confidentiality and privacy requirements of dataset owners and model owners, outperforming the state-of-the-art privacy-preserving training systems by up to 543x on CPU and 113x on GPU TEEs.
☆ Spend More to Save More (SM2): An Energy-Aware Implementation of Successive Halving for Sustainable Hyperparameter Optimization
A fundamental step in the development of machine learning models commonly involves the tuning of hyperparameters, often leading to multiple model training runs to work out the best-performing configuration. As machine learning tasks and models grow in complexity, there is an escalating need for solutions that not only improve performance but also address sustainability concerns. Existing strategies predominantly focus on maximizing the performance of the model without considering energy efficiency. To bridge this gap, in this paper, we introduce Spend More to Save More (SM2), an energy-aware hyperparameter optimization implementation based on the widely adopted successive halving algorithm. Unlike conventional approaches including energy-intensive testing of individual hyperparameter configurations, SM2 employs exploratory pretraining to identify inefficient configurations with minimal energy expenditure. Incorporating hardware characteristics and real-time energy consumption tracking, SM2 identifies an optimal configuration that not only maximizes the performance of the model but also enables energy-efficient training. Experimental validations across various datasets, models, and hardware setups confirm the efficacy of SM2 to prevent the waste of energy during the training of hyperparameter configurations.
☆ Enhancing Interpretability Through Loss-Defined Classification Objective in Structured Latent Spaces
Supervised machine learning often operates on the data-driven paradigm, wherein internal model parameters are autonomously optimized to converge predicted outputs with the ground truth, devoid of explicitly programming rules or a priori assumptions. Although data-driven methods have yielded notable successes across various benchmark datasets, they inherently treat models as opaque entities, thereby limiting their interpretability and yielding a lack of explanatory insights into their decision-making processes. In this work, we introduce Latent Boost, a novel approach that integrates advanced distance metric learning into supervised classification tasks, enhancing both interpretability and training efficiency. Thus during training, the model is not only optimized for classification metrics of the discrete data points but also adheres to the rule that the collective representation zones of each class should be sharply clustered. By leveraging the rich structural insights of intermediate model layer latent representations, Latent Boost improves classification interpretability, as demonstrated by higher Silhouette scores, while accelerating training convergence. These performance and latent structural benefits are achieved with minimum additional cost, making it broadly applicable across various datasets without requiring data-specific adjustments. Furthermore, Latent Boost introduces a new paradigm for aligning classification performance with improved model transparency to address the challenges of black-box models.
☆ Image-Based Malware Classification Using QR and Aztec Codes
In recent years, the use of image-based techniques for malware detection has gained prominence, with numerous studies demonstrating the efficacy of deep learning approaches such as Convolutional Neural Networks (CNN) in classifying images derived from executable files. In this paper, we consider an innovative method that relies on an image conversion process that consists of transforming features extracted from executable files into QR and Aztec codes. These codes capture structural patterns in a format that may enhance the learning capabilities of CNNs. We design and implement CNN architectures tailored to the unique properties of these codes and apply them to a comprehensive analysis involving two extensive malware datasets, both of which include a significant corpus of benign samples. Our results yield a split decision, with CNNs trained on QR and Aztec codes outperforming the state of the art on one of the datasets, but underperforming more typical techniques on the other dataset. These results indicate that the use of QR and Aztec codes as a form of feature engineering holds considerable promise in the malware domain, and that additional research is needed to better understand the relative strengths and weaknesses of such an approach.
☆ REPEAT: Improving Uncertainty Estimation in Representation Learning Explainability AAAI 2025
Incorporating uncertainty is crucial to provide trustworthy explanations of deep learning models. Recent works have demonstrated how uncertainty modeling can be particularly important in the unsupervised field of representation learning explainable artificial intelligence (R-XAI). Current R-XAI methods provide uncertainty by measuring variability in the importance score. However, they fail to provide meaningful estimates of whether a pixel is certainly important or not. In this work, we propose a new R-XAI method called REPEAT that addresses the key question of whether or not a pixel is \textit{certainly} important. REPEAT leverages the stochasticity of current R-XAI methods to produce multiple estimates of importance, thus considering each pixel in an image as a Bernoulli random variable that is either important or unimportant. From these Bernoulli random variables we can directly estimate the importance of a pixel and its associated certainty, thus enabling users to determine certainty in pixel importance. Our extensive evaluation shows that REPEAT gives certainty estimates that are more intuitive, better at detecting out-of-distribution data, and more concise.
comment: Accepted at AAAI 2025. Code available at: https://github.com/Wickstrom/REPEAT
☆ GradStop: Exploring Training Dynamics in Unsupervised Outlier Detection through Gradient Cohesion
Unsupervised Outlier Detection (UOD) is a critical task in data mining and machine learning, aiming to identify instances that significantly deviate from the majority. Without any label, deep UOD methods struggle with the misalignment between the model's direct optimization goal and the final performance goal of Outlier Detection (OD) task. Through the perspective of training dynamics, this paper proposes an early stopping algorithm to optimize the training of deep UOD models, ensuring they perform optimally in OD rather than overfitting the entire contaminated dataset. Inspired by UOD mechanism and inlier priority phenomenon, where intuitively models fit inliers more quickly than outliers, we propose GradStop, a sampling-based label-free algorithm to estimate model's real-time performance during training. First, a sampling method generates two sets: one likely containing more outliers and the other more inliers, then a metric based on gradient cohesion is applied to probe into current training dynamics, which reflects model's performance on OD task. Experimental results on 4 deep UOD algorithms and 47 real-world datasets and theoretical proofs demonstrate the effectiveness of our proposed early stopping algorithm in enhancing the performance of deep UOD models. Auto Encoder (AE) enhanced by GradStop achieves better performance than itself, other SOTA UOD methods, and even ensemble AEs. Our method provides a robust and effective solution to the problem of performance degradation during training, enabling deep UOD models to achieve better potential in anomaly detection tasks.
☆ InvDiff: Invariant Guidance for Bias Mitigation in Diffusion Models KDD 2025
As one of the most successful generative models, diffusion models have demonstrated remarkable efficacy in synthesizing high-quality images. These models learn the underlying high-dimensional data distribution in an unsupervised manner. Despite their success, diffusion models are highly data-driven and prone to inheriting the imbalances and biases present in real-world data. Some studies have attempted to address these issues by designing text prompts for known biases or using bias labels to construct unbiased data. While these methods have shown improved results, real-world scenarios often contain various unknown biases, and obtaining bias labels is particularly challenging. In this paper, we emphasize the necessity of mitigating bias in pre-trained diffusion models without relying on auxiliary bias annotations. To tackle this problem, we propose a framework, InvDiff, which aims to learn invariant semantic information for diffusion guidance. Specifically, we propose identifying underlying biases in the training data and designing a novel debiasing training objective. Then, we employ a lightweight trainable module that automatically preserves invariant semantic information and uses it to guide the diffusion model's sampling process toward unbiased outcomes simultaneously. Notably, we only need to learn a small number of parameters in the lightweight learnable module without altering the pre-trained diffusion model. Furthermore, we provide a theoretical guarantee that the implementation of InvDiff is equivalent to reducing the error upper bound of generalization. Extensive experimental results on three publicly available benchmarks demonstrate that InvDiff effectively reduces biases while maintaining the quality of image generation. Our code is available at https://github.com/Hundredl/InvDiff.
comment: KDD 2025
☆ IRL for Restless Multi-Armed Bandits with Applications in Maternal and Child Health
Public health practitioners often have the goal of monitoring patients and maximizing patients' time spent in "favorable" or healthy states while being constrained to using limited resources. Restless multi-armed bandits (RMAB) are an effective model to solve this problem as they are helpful to allocate limited resources among many agents under resource constraints, where patients behave differently depending on whether they are intervened on or not. However, RMABs assume the reward function is known. This is unrealistic in many public health settings because patients face unique challenges and it is impossible for a human to know who is most deserving of any intervention at such a large scale. To address this shortcoming, this paper is the first to present the use of inverse reinforcement learning (IRL) to learn desired rewards for RMABs, and we demonstrate improved outcomes in a maternal and child health telehealth program. First we allow public health experts to specify their goals at an aggregate or population level and propose an algorithm to design expert trajectories at scale based on those goals. Second, our algorithm WHIRL uses gradient updates to optimize the objective, allowing for efficient and accurate learning of RMAB rewards. Third, we compare with existing baselines and outperform those in terms of run-time and accuracy. Finally, we evaluate and show the usefulness of WHIRL on thousands on beneficiaries from a real-world maternal and child health setting in India. We publicly release our code here: https://github.com/Gjain234/WHIRL.
☆ Federated Learning for Traffic Flow Prediction with Synthetic Data Augmentation
Deep-learning based traffic prediction models require vast amounts of data to learn embedded spatial and temporal dependencies. The inherent privacy and commercial sensitivity of such data has encouraged a shift towards decentralised data-driven methods, such as Federated Learning (FL). Under a traditional Machine Learning paradigm, traffic flow prediction models can capture spatial and temporal relationships within centralised data. In reality, traffic data is likely distributed across separate data silos owned by multiple stakeholders. In this work, a cross-silo FL setting is motivated to facilitate stakeholder collaboration for optimal traffic flow prediction applications. This work introduces an FL framework, referred to as FedTPS, to generate synthetic data to augment each client's local dataset by training a diffusion-based trajectory generation model through FL. The proposed framework is evaluated on a large-scale real world ride-sharing dataset using various FL methods and Traffic Flow Prediction models, including a novel prediction model we introduce, which leverages Temporal and Graph Attention mechanisms to learn the Spatio-Temporal dependencies embedded within regional traffic flow data. Experimental results show that FedTPS outperforms multiple other FL baselines with respect to global model performance.
comment: 11 pages, 7 figures, 6 tables, ACM format
☆ Efficient Rectification of Neuro-Symbolic Reasoning Inconsistencies by Abductive Reflection AAAI 2025
Neuro-Symbolic (NeSy) AI could be regarded as an analogy to human dual-process cognition, modeling the intuitive System 1 with neural networks and the algorithmic System 2 with symbolic reasoning. However, for complex learning targets, NeSy systems often generate outputs inconsistent with domain knowledge and it is challenging to rectify them. Inspired by the human Cognitive Reflection, which promptly detects errors in our intuitive response and revises them by invoking the System 2 reasoning, we propose to improve NeSy systems by introducing Abductive Reflection (ABL-Refl) based on the Abductive Learning (ABL) framework. ABL-Refl leverages domain knowledge to abduce a reflection vector during training, which can then flag potential errors in the neural network outputs and invoke abduction to rectify them and generate consistent outputs during inference. ABL-Refl is highly efficient in contrast to previous ABL implementations. Experiments show that ABL-Refl outperforms state-of-the-art NeSy methods, achieving excellent accuracy with fewer training resources and enhanced efficiency.
comment: Accepted to AAAI 2025
☆ On best approximation by multivariate ridge functions with applications to generalized translation networks
We prove sharp upper and lower bounds for the approximation of Sobolev functions by sums of multivariate ridge functions, i.e., functions of the form $\mathbb{R}^d \ni x \mapsto \sum_{k=1}^n h_k(A_k x) \in \mathbb{R}$ with $h_k : \mathbb{R}^\ell \to \mathbb{R}$ and $A_k \in \mathbb{R}^{\ell \times d}$. We show that the order of approximation asymptotically behaves as $n^{-r/(d-\ell)}$, where $r$ is the regularity of the Sobolev functions to be approximated. Our lower bound even holds when approximating $L^\infty$-Sobolev functions of regularity $r$ with error measured in $L^1$, while our upper bound applies to the approximation of $L^p$-Sobolev functions in $L^p$ for any $1 \leq p \leq \infty$. These bounds generalize well-known results about the approximation properties of univariate ridge functions to the multivariate case. Moreover, we use these bounds to obtain sharp asymptotic bounds for the approximation of Sobolev functions using generalized translation networks and complex-valued neural networks.
☆ From Multimodal LLMs to Generalist Embodied Agents: Methods and Lessons
We examine the capability of Multimodal Large Language Models (MLLMs) to tackle diverse domains that extend beyond the traditional language and vision tasks these models are typically trained on. Specifically, our focus lies in areas such as Embodied AI, Games, UI Control, and Planning. To this end, we introduce a process of adapting an MLLM to a Generalist Embodied Agent (GEA). GEA is a single unified model capable of grounding itself across these varied domains through a multi-embodiment action tokenizer. GEA is trained with supervised learning on a large dataset of embodied experiences and with online RL in interactive simulators. We explore the data and algorithmic choices necessary to develop such a model. Our findings reveal the importance of training with cross-domain data and online RL for building generalist agents. The final GEA model achieves strong generalization performance to unseen tasks across diverse benchmarks compared to other generalist models and benchmark-specific approaches.
☆ Proactive Model Adaptation Against Concept Drift for Online Time Series Forecasting KDD 2025
Time series forecasting always faces the challenge of concept drift, where data distributions evolve over time, leading to a decline in forecast model performance. Existing solutions are based on online learning, which continually organize recent time series observations as new training samples and update model parameters according to the forecasting feedback on recent data. However, they overlook a critical issue: obtaining ground-truth future values of each sample should be delayed until after the forecast horizon. This delay creates a temporal gap between the training samples and the test sample. Our empirical analysis reveals that the gap can introduce concept drift, causing forecast models to adapt to outdated concepts. In this paper, we present \textsc{Proceed}, a novel proactive model adaptation framework for online time series forecasting. \textsc{Proceed} first operates by estimating the concept drift between the recently used training samples and the current test sample. It then employs an adaptation generator to efficiently translate the estimated drift into parameter adjustments, proactively adapting the model to the test sample. To enhance the generalization capability of the framework, \textsc{Proceed} is trained on synthetic diverse concept drifts. We conduct extensive experiments on five real-world datasets across various forecast models. The empirical study demonstrates that our proposed \textsc{Proceed} brings more performance improvements than the state-of-the-art online learning methods, significantly facilitating forecast models' resilience against concept drifts.
comment: Accepted by KDD 2025. Preprint version
☆ Koopman Theory-Inspired Method for Learning Time Advancement Operators in Unstable Flame Front Evolution
Predicting the evolution of complex systems governed by partial differential equations (PDEs) remains challenging, especially for nonlinear, chaotic behaviors. This study introduces Koopman-inspired Fourier Neural Operators (kFNO) and Convolutional Neural Networks (kCNN) to learn solution advancement operators for flame front instabilities. By transforming data into a high-dimensional latent space, these models achieve more accurate multi-step predictions compared to traditional methods. Benchmarking across one- and two-dimensional flame front scenarios demonstrates the proposed approaches' superior performance in short-term accuracy and long-term statistical reproduction, offering a promising framework for modeling complex dynamical systems.
comment: 28 pages, 12 figures
☆ From Logistic Regression to the Perceptron Algorithm: Exploring Gradient Descent with Large Step Sizes
We focus on the classification problem with a separable dataset, one of the most important and classical problems from machine learning. The standard approach to this task is logistic regression with gradient descent (LR+GD). Recent studies have observed that LR+GD can find a solution with arbitrarily large step sizes, defying conventional optimization theory. Our work investigates this phenomenon and makes three interconnected key observations about LR+GD with large step sizes. First, we find a remarkably simple explanation of why LR+GD with large step sizes solves the classification problem: LR+GD reduces to a batch version of the celebrated perceptron algorithm when the step size $\gamma \to \infty.$ Second, we observe that larger step sizes lead LR+GD to higher logistic losses when it tends to the perceptron algorithm, but larger step sizes also lead to faster convergence to a solution for the classification problem, meaning that logistic loss is an unreliable metric of the proximity to a solution. Surprisingly, high loss values can actually indicate faster convergence. Third, since the convergence rate in terms of loss function values of LR+GD is unreliable, we examine the iteration complexity required by LR+GD with large step sizes to solve the classification problem and prove that this complexity is suboptimal. To address this, we propose a new method, Normalized LR+GD - based on the connection between LR+GD and the perceptron algorithm - with much better theoretical guarantees.
☆ Robustness of Graph Classification: failure modes, causes, and noise-resistant loss in Graph Neural Networks
Graph Neural Networks (GNNs) are powerful at solving graph classification tasks, yet applied problems often contain noisy labels. In this work, we study GNN robustness to label noise, demonstrate GNN failure modes when models struggle to generalise on low-order graphs, low label coverage, or when a model is over-parameterized. We establish both empirical and theoretical links between GNN robustness and the reduction of the total Dirichlet Energy of learned node representations, which encapsulates the hypothesized GNN smoothness inductive bias. Finally, we introduce two training strategies to enhance GNN robustness: (1) by incorporating a novel inductive bias in the weight matrices through the removal of negative eigenvalues, connected to Dirichlet Energy minimization; (2) by extending to GNNs a loss penalty that promotes learned smoothness. Importantly, neither approach negatively impacts performance in noise-free settings, supporting our hypothesis that the source of GNNs robustness is their smoothness inductive bias.
☆ Grasp Diffusion Network: Learning Grasp Generators from Partial Point Clouds with Diffusion Models in SO(3)xR3
Grasping objects successfully from a single-view camera is crucial in many robot manipulation tasks. An approach to solve this problem is to leverage simulation to create large datasets of pairs of objects and grasp poses, and then learn a conditional generative model that can be prompted quickly during deployment. However, the grasp pose data is highly multimodal since there are several ways to grasp an object. Hence, in this work, we learn a grasp generative model with diffusion models to sample candidate grasp poses given a partial point cloud of an object. A novel aspect of our method is to consider diffusion in the manifold space of rotations and to propose a collision-avoidance cost guidance to improve the grasp success rate during inference. To accelerate grasp sampling we use recent techniques from the diffusion literature to achieve faster inference times. We show in simulation and real-world experiments that our approach can grasp several objects from raw depth images with $90\%$ success rate and benchmark it against several baselines.
☆ Adversarial Purification by Consistency-aware Latent Space Optimization on Data Manifolds
Deep neural networks (DNNs) are vulnerable to adversarial samples crafted by adding imperceptible perturbations to clean data, potentially leading to incorrect and dangerous predictions. Adversarial purification has been an effective means to improve DNNs robustness by removing these perturbations before feeding the data into the model. However, it faces significant challenges in preserving key structural and semantic information of data, as the imperceptible nature of adversarial perturbations makes it hard to avoid over-correcting, which can destroy important information and degrade model performance. In this paper, we break away from traditional adversarial purification methods by focusing on the clean data manifold. To this end, we reveal that samples generated by a well-trained generative model are close to clean ones but far from adversarial ones. Leveraging this insight, we propose Consistency Model-based Adversarial Purification (CMAP), which optimizes vectors within the latent space of a pre-trained consistency model to generate samples for restoring clean data. Specifically, 1) we propose a \textit{Perceptual consistency restoration} mechanism by minimizing the discrepancy between generated samples and input samples in both pixel and perceptual spaces. 2) To maintain the optimized latent vectors within the valid data manifold, we introduce a \textit{Latent distribution consistency constraint} strategy to align generated samples with the clean data distribution. 3) We also apply a \textit{Latent vector consistency prediction} scheme via an ensemble approach to enhance prediction reliability. CMAP fundamentally addresses adversarial perturbations at their source, providing a robust purification. Extensive experiments on CIFAR-10 and ImageNet-100 show that our CMAP significantly enhances robustness against strong adversarial attacks while preserving high natural accuracy.
comment: 17 pages, 8 figures
☆ NyayaAnumana & INLegalLlama: The Largest Indian Legal Judgment Prediction Dataset and Specialized Language Model for Enhanced Decision Analysis COLING 2025
The integration of artificial intelligence (AI) in legal judgment prediction (LJP) has the potential to transform the legal landscape, particularly in jurisdictions like India, where a significant backlog of cases burdens the legal system. This paper introduces NyayaAnumana, the largest and most diverse corpus of Indian legal cases compiled for LJP, encompassing a total of 7,02,945 preprocessed cases. NyayaAnumana, which combines the words "Nyay" (judgment) and "Anuman" (prediction or inference) respectively for most major Indian languages, includes a wide range of cases from the Supreme Court, High Courts, Tribunal Courts, District Courts, and Daily Orders and, thus, provides unparalleled diversity and coverage. Our dataset surpasses existing datasets like PredEx and ILDC, offering a comprehensive foundation for advanced AI research in the legal domain. In addition to the dataset, we present INLegalLlama, a domain-specific generative large language model (LLM) tailored to the intricacies of the Indian legal system. It is developed through a two-phase training approach over a base LLaMa model. First, Indian legal documents are injected using continual pretraining. Second, task-specific supervised finetuning is done. This method allows the model to achieve a deeper understanding of legal contexts. Our experiments demonstrate that incorporating diverse court data significantly boosts model accuracy, achieving approximately 90% F1-score in prediction tasks. INLegalLlama not only improves prediction accuracy but also offers comprehensible explanations, addressing the need for explainability in AI-assisted legal decisions.
comment: Accepted on COLING 2025
☆ Backdoor attacks on DNN and GBDT -- A Case Study from the insurance domain
Machine learning (ML) will likely play a large role in many processes in the future, also for insurance companies. However, ML models are at risk of being attacked and manipulated. In this work, the robustness of Gradient Boosted Decision Tree (GBDT) models and Deep Neural Networks (DNN) within an insurance context will be evaluated. Therefore, two GBDT models and two DNNs are trained on two different tabular datasets from an insurance context. Past research in this domain mainly used homogenous data and there are comparably few insights regarding heterogenous tabular data. The ML tasks performed on the datasets are claim prediction (regression) and fraud detection (binary classification). For the backdoor attacks different samples containing a specific pattern were crafted and added to the training data. It is shown, that this type of attack can be highly successful, even with a few added samples. The backdoor attacks worked well on the models trained on one dataset but poorly on the models trained on the other. In real-world scenarios the attacker will have to face several obstacles but as attacks can work with very few added samples this risk should be evaluated.
comment: 40 pages, 14 figures
☆ Zero-Shot Mono-to-Binaural Speech Synthesis
We present ZeroBAS, a neural method to synthesize binaural audio from monaural audio recordings and positional information without training on any binaural data. To our knowledge, this is the first published zero-shot neural approach to mono-to-binaural audio synthesis. Specifically, we show that a parameter-free geometric time warping and amplitude scaling based on source location suffices to get an initial binaural synthesis that can be refined by iteratively applying a pretrained denoising vocoder. Furthermore, we find this leads to generalization across room conditions, which we measure by introducing a new dataset, TUT Mono-to-Binaural, to evaluate state-of-the-art monaural-to-binaural synthesis methods on unseen conditions. Our zero-shot method is perceptually on-par with the performance of supervised methods on the standard mono-to-binaural dataset, and even surpasses them on our out-of-distribution TUT Mono-to-Binaural dataset. Our results highlight the potential of pretrained generative audio models and zero-shot learning to unlock robust binaural audio synthesis.
☆ Benchmarking learned algorithms for computed tomography image reconstruction tasks
Computed tomography (CT) is a widely used non-invasive diagnostic method in various fields, and recent advances in deep learning have led to significant progress in CT image reconstruction. However, the lack of large-scale, open-access datasets has hindered the comparison of different types of learned methods. To address this gap, we use the 2DeteCT dataset, a real-world experimental computed tomography dataset, for benchmarking machine learning based CT image reconstruction algorithms. We categorize these methods into post-processing networks, learned/unrolled iterative methods, learned regularizer methods, and plug-and-play methods, and provide a pipeline for easy implementation and evaluation. Using key performance metrics, including SSIM and PSNR, our benchmarking results showcase the effectiveness of various algorithms on tasks such as full data reconstruction, limited-angle reconstruction, sparse-angle reconstruction, low-dose reconstruction, and beam-hardening corrected reconstruction. With this benchmarking study, we provide an evaluation of a range of algorithms representative for different categories of learned reconstruction methods on a recently published dataset of real-world experimental CT measurements. The reproducible setup of methods and CT image reconstruction tasks in an open-source toolbox enables straightforward addition and comparison of new methods later on. The toolbox also provides the option to load the 2DeteCT dataset differently for extensions to other problems and different CT reconstruction tasks.
☆ ALoRE: Efficient Visual Adaptation via Aggregating Low Rank Experts
Parameter-efficient transfer learning (PETL) has become a promising paradigm for adapting large-scale vision foundation models to downstream tasks. Typical methods primarily leverage the intrinsic low rank property to make decomposition, learning task-specific weights while compressing parameter size. However, such approaches predominantly manipulate within the original feature space utilizing a single-branch structure, which might be suboptimal for decoupling the learned representations and patterns. In this paper, we propose ALoRE, a novel PETL method that reuses the hypercomplex parameterized space constructed by Kronecker product to Aggregate Low Rank Experts using a multi-branch paradigm, disentangling the learned cognitive patterns during training. Thanks to the artful design, ALoRE maintains negligible extra parameters and can be effortlessly merged into the frozen backbone via re-parameterization in a sequential manner, avoiding additional inference latency. We conduct extensive experiments on 24 image classification tasks using various backbone variants. Experimental results demonstrate that ALoRE outperforms the full fine-tuning strategy and other state-of-the-art PETL methods in terms of performance and parameter efficiency. For instance, ALoRE obtains 3.06% and 9.97% Top-1 accuracy improvement on average compared to full fine-tuning on the FGVC datasets and VTAB-1k benchmark by only updating 0.15M parameters.
comment: 23 pages, 7 figures
☆ Post-Hoc MOTS: Exploring the Capabilities of Time-Symmetric Multi-Object Tracking
Temporal forward-tracking has been the dominant approach for multi-object segmentation and tracking (MOTS). However, a novel time-symmetric tracking methodology has recently been introduced for the detection, segmentation, and tracking of budding yeast cells in pre-recorded samples. Although this architecture has demonstrated a unique perspective on stable and consistent tracking, as well as missed instance re-interpolation, its evaluation has so far been largely confined to settings related to videomicroscopic environments. In this work, we aim to reveal the broader capabilities, advantages, and potential challenges of this architecture across various specifically designed scenarios, including a pedestrian tracking dataset. We also conduct an ablation study comparing the model against its restricted variants and the widely used Kalman filter. Furthermore, we present an attention analysis of the tracking architecture for both pretrained and non-pretrained models
☆ A Unified Model For Voice and Accent Conversion In Speech and Singing using Self-Supervised Learning and Feature Extraction
This paper presents a new voice conversion model capable of transforming both speaking and singing voices. It addresses key challenges in current systems, such as conveying emotions, managing pronunciation and accent changes, and reproducing non-verbal sounds. One of the model's standout features is its ability to perform accent conversion on hybrid voice samples that encompass both speech and singing, allowing it to change the speaker's accent while preserving the original content and prosody. The proposed model uses an encoder-decoder architecture: the encoder is based on HuBERT to process the speech's acoustic and linguistic content, while the HiFi-GAN decoder audio matches the target speaker's voice. The model incorporates fundamental frequency (f0) features and singer embeddings to enhance performance while ensuring the pitch & tone accuracy and vocal identity are preserved during transformation. This approach improves how naturally and flexibly voice style can be transformed, showing strong potential for applications in voice dubbing, content creation, and technologies like Text-to-Speech (TTS) and Interactive Voice Response (IVR) systems.
comment: 7 pages, 5 figures, 2 tables
☆ Edge-Splitting MLP: Node Classification on Homophilic and Heterophilic Graphs without Message Passing
Message Passing Neural Networks (MPNNs) have demonstrated remarkable success in node classification on homophilic graphs. It has been shown that they do not solely rely on homophily but on neighborhood distributions of nodes, i.e., consistency of the neighborhood label distribution within the same class. MLP-based models do not use message passing, \eg Graph-MLP incorporates the neighborhood in a separate loss function. These models are faster and more robust to edge noise. Graph-MLP maps adjacent nodes closer in the embedding space but is unaware of the neighborhood pattern of the labels, i.e., relies solely on homophily. Edge Splitting GNN (ES-GNN) is a model specialized for heterophilic graphs and splits the edges into task-relevant and task-irrelevant, respectively. To mitigate the limitations of Graph-MLP on heterophilic graphs, we propose ES-MLP that combines Graph-MLP with an edge-splitting mechanism from ES-GNN. It incorporates the edge splitting into the loss of Graph-MLP to learn two separate adjacency matrices based on relevant and irrelevant feature pairs. Our experiments on seven datasets with six baselines show that ES-MLP is on par with homophilic and heterophilic models on all datasets without using edges during inference. We show that ES-MLP is robust to multiple types of edge noise during inference and that its inference time is two to five times faster than that of commonly used MPNNs. The source code is available at https://github.com/MatthiasKohn/ES-MLP.
comment: Published at Learning on Graphs, 2024
☆ Enhancing Cybersecurity in IoT Networks: A Deep Learning Approach to Anomaly Detection
With the proliferation of the Internet and smart devices, IoT technology has seen significant advancements and has become an integral component of smart homes, urban security, smart logistics, and other sectors. IoT facilitates real-time monitoring of critical production indicators, enabling businesses to detect potential quality issues, anticipate equipment malfunctions, and refine processes, thereby minimizing losses and reducing costs. Furthermore, IoT enhances real-time asset tracking, optimizing asset utilization and management. However, the expansion of IoT has also led to a rise in cybercrimes, with devices increasingly serving as vectors for malicious attacks. As the number of IoT devices grows, there is an urgent need for robust network security measures to counter these escalating threats. This paper introduces a deep learning model incorporating LSTM and attention mechanisms, a pivotal strategy in combating cybercrime in IoT networks. Our experiments, conducted on datasets including IoT-23, BoT-IoT, IoT network intrusion, MQTT, and MQTTset, demonstrate that our proposed method outperforms existing baselines.
☆ GDSG: Graph Diffusion-based Solution Generation for Optimization Problems in MEC Networks
Optimization is crucial for MEC networks to function efficiently and reliably, most of which are NP-hard and lack efficient approximation algorithms. This leads to a paucity of optimal solution, constraining the effectiveness of conventional deep learning approaches. Most existing learning-based methods necessitate extensive optimal data and fail to exploit the potential benefits of suboptimal data that can be obtained with greater efficiency and effectiveness. Taking the multi-server multi-user computation offloading (MSCO) problem, which is widely observed in systems like Internet-of-Vehicles (IoV) and Unmanned Aerial Vehicle (UAV) networks, as a concrete scenario, we present a Graph Diffusion-based Solution Generation (GDSG) method. This approach is designed to work with suboptimal datasets while converging to the optimal solution large probably. We transform the optimization issue into distribution-learning and offer a clear explanation of learning from suboptimal training datasets. We build GDSG as a multi-task diffusion model utilizing a Graph Neural Network (GNN) to acquire the distribution of high-quality solutions. We use a simple and efficient heuristic approach to obtain a sufficient amount of training data composed entirely of suboptimal solutions. In our implementation, we enhance the backbone GNN and achieve improved generalization. GDSG also reaches nearly 100\% task orthogonality, ensuring no interference between the discrete and continuous generation tasks. We further reveal that this orthogonality arises from the diffusion-related training loss, rather than the neural network architecture itself. The experiments demonstrate that GDSG surpasses other benchmark methods on both the optimal and suboptimal training datasets. The MSCO datasets has open-sourced at http://ieee-dataport.org/13824, as well as the GDSG algorithm codes at https://github.com/qiyu3816/GDSG.
☆ SINERGYM -- A virtual testbed for building energy optimization with Reinforcement Learning
Simulation has become a crucial tool for Building Energy Optimization (BEO) as it enables the evaluation of different design and control strategies at a low cost. Machine Learning (ML) algorithms can leverage large-scale simulations to learn optimal control from vast amounts of data without supervision, particularly under the Reinforcement Learning (RL) paradigm. Unfortunately, the lack of open and standardized tools has hindered the widespread application of ML and RL to BEO. To address this issue, this paper presents Sinergym, an open-source Python-based virtual testbed for large-scale building simulation, data collection, continuous control, and experiment monitoring. Sinergym provides a consistent interface for training and running controllers, predefined benchmarks, experiment visualization and replication support, and comprehensive documentation in a ready-to-use software library. This paper 1) highlights the main features of Sinergym in comparison to other existing frameworks, 2) describes its basic usage, and 3) demonstrates its applicability for RL-based BEO through several representative examples. By integrating simulation, data, and control, Sinergym supports the development of intelligent, data-driven applications for more efficient and responsive building operations, aligning with the objectives of digital twin technology.
☆ Self-Refining Diffusion Samplers: Enabling Parallelization via Parareal Iterations NeurIPS 2024
In diffusion models, samples are generated through an iterative refinement process, requiring hundreds of sequential model evaluations. Several recent methods have introduced approximations (fewer discretization steps or distillation) to trade off speed at the cost of sample quality. In contrast, we introduce Self-Refining Diffusion Samplers (SRDS) that retain sample quality and can improve latency at the cost of additional parallel compute. We take inspiration from the Parareal algorithm, a popular numerical method for parallel-in-time integration of differential equations. In SRDS, a quick but rough estimate of a sample is first created and then iteratively refined in parallel through Parareal iterations. SRDS is not only guaranteed to accurately solve the ODE and converge to the serial solution but also benefits from parallelization across the diffusion trajectory, enabling batched inference and pipelining. As we demonstrate for pre-trained diffusion models, the early convergence of this refinement procedure drastically reduces the number of steps required to produce a sample, speeding up generation for instance by up to 1.7x on a 25-step StableDiffusion-v2 benchmark and up to 4.3x on longer trajectories.
comment: NeurIPS 2024
☆ k-HyperEdge Medoids for Clustering Ensemble
Clustering ensemble has been a popular research topic in data science due to its ability to improve the robustness of the single clustering method. Many clustering ensemble methods have been proposed, most of which can be categorized into clustering-view and sample-view methods. The clustering-view method is generally efficient, but it could be affected by the unreliability that existed in base clustering results. The sample-view method shows good performance, while the construction of the pairwise sample relation is time-consuming. In this paper, the clustering ensemble is formulated as a k-HyperEdge Medoids discovery problem and a clustering ensemble method based on k-HyperEdge Medoids that considers the characteristics of the above two types of clustering ensemble methods is proposed. In the method, a set of hyperedges is selected from the clustering view efficiently, then the hyperedges are diffused and adjusted from the sample view guided by a hyperedge loss function to construct an effective k-HyperEdge Medoid set. The loss function is mainly reduced by assigning samples to the hyperedge with the highest degree of belonging. Theoretical analyses show that the solution can approximate the optimal, the assignment method can gradually reduce the loss function, and the estimation of the belonging degree is statistically reasonable. Experiments on artificial data show the working mechanism of the proposed method. The convergence of the method is verified by experimental analysis of twenty data sets. The effectiveness and efficiency of the proposed method are also verified on these data, with nine representative clustering ensemble algorithms as reference.
☆ DistrictNet: Decision-aware learning for geographical districting NeurIPS 2024
Districting is a complex combinatorial problem that consists in partitioning a geographical area into small districts. In logistics, it is a major strategic decision determining operating costs for several years. Solving districting problems using traditional methods is intractable even for small geographical areas and existing heuristics often provide sub-optimal results. We present a structured learning approach to find high-quality solutions to real-world districting problems in a few minutes. It is based on integrating a combinatorial optimization layer, the capacitated minimum spanning tree problem, into a graph neural network architecture. To train this pipeline in a decision-aware fashion, we show how to construct target solutions embedded in a suitable space and learn from target solutions. Experiments show that our approach outperforms existing methods as it can significantly reduce costs on real-world cities.
comment: Accepted at NeurIPS 2024
☆ Towards Precision in Bolted Joint Design: A Preliminary Machine Learning-Based Parameter Prediction
Bolted joints are critical in engineering for maintaining structural integrity and reliability. Accurate prediction of parameters influencing their function and behavior is essential for optimal performance. Traditional methods often fail to capture the non-linear behavior of bolted joints or require significant computational resources, limiting accuracy and efficiency. This study addresses these limitations by combining empirical data with a feed-forward neural network to predict load capacity and friction coefficients. Leveraging experimental data and systematic preprocessing, the model effectively captures nonlinear relationships, including rescaling output variables to address scale discrepancies, achieving 95.24% predictive accuracy. While limited dataset size and diversity restrict generalizability, the findings demonstrate the potential of neural networks as a reliable, efficient alternative for bolted joint design. Future work will focus on expanding datasets and exploring hybrid modeling techniques to enhance applicability.
comment: 10 pages, 6 figures, submitted to: ICED25 - 25th International Conference on Engineering Design, in Review
☆ Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective AAAI 2025
To address catastrophic forgetting in Continual Relation Extraction (CRE), many current approaches rely on memory buffers to rehearse previously learned knowledge while acquiring new tasks. Recently, prompt-based methods have emerged as potent alternatives to rehearsal-based strategies, demonstrating strong empirical performance. However, upon analyzing existing prompt-based approaches for CRE, we identified several critical limitations, such as inaccurate prompt selection, inadequate mechanisms for mitigating forgetting in shared parameters, and suboptimal handling of cross-task and within-task variances. To overcome these challenges, we draw inspiration from the relationship between prefix-tuning and mixture of experts, proposing a novel approach that employs a prompt pool for each task, capturing variations within each task while enhancing cross-task variances. Furthermore, we incorporate a generative model to consolidate prior knowledge within shared parameters, eliminating the need for explicit data storage. Extensive experiments validate the efficacy of our approach, demonstrating superior performance over state-of-the-art prompt-based and rehearsal-free methods in continual relation extraction.
comment: Accepted to AAAI 2025
☆ How Does the Smoothness Approximation Method Facilitate Generalization for Federated Adversarial Learning?
Federated Adversarial Learning (FAL) is a robust framework for resisting adversarial attacks on federated learning. Although some FAL studies have developed efficient algorithms, they primarily focus on convergence performance and overlook generalization. Generalization is crucial for evaluating algorithm performance on unseen data. However, generalization analysis is more challenging due to non-smooth adversarial loss functions. A common approach to addressing this issue is to leverage smoothness approximation. In this paper, we develop algorithm stability measures to evaluate the generalization performance of two popular FAL algorithms: \textit{Vanilla FAL (VFAL)} and {\it Slack FAL (SFAL)}, using three different smooth approximation methods: 1) \textit{Surrogate Smoothness Approximation (SSA)}, (2) \textit{Randomized Smoothness Approximation (RSA)}, and (3) \textit{Over-Parameterized Smoothness Approximation (OPSA)}. Based on our in-depth analysis, we answer the question of how to properly set the smoothness approximation method to mitigate generalization error in FAL. Moreover, we identify RSA as the most effective method for reducing generalization error. In highly data-heterogeneous scenarios, we also recommend employing SFAL to mitigate the deterioration of generalization performance caused by heterogeneity. Based on our theoretical results, we provide insights to help develop more efficient FAL algorithms, such as designing new metrics and dynamic aggregation rules to mitigate heterogeneity.
☆ FLIP: Flow-Centric Generative Planning for General-Purpose Manipulation Tasks
We aim to develop a model-based planning framework for world models that can be scaled with increasing model and data budgets for general-purpose manipulation tasks with only language and vision inputs. To this end, we present FLow-centric generative Planning (FLIP), a model-based planning algorithm on visual space that features three key modules: 1. a multi-modal flow generation model as the general-purpose action proposal module; 2. a flow-conditioned video generation model as the dynamics module; and 3. a vision-language representation learning model as the value module. Given an initial image and language instruction as the goal, FLIP can progressively search for long-horizon flow and video plans that maximize the discounted return to accomplish the task. FLIP is able to synthesize long-horizon plans across objects, robots, and tasks with image flows as the general action representation, and the dense flow information also provides rich guidance for long-horizon video generation. In addition, the synthesized flow and video plans can guide the training of low-level control policies for robot execution. Experiments on diverse benchmarks demonstrate that FLIP can improve both the success rates and quality of long-horizon video plan synthesis and has the interactive world model property, opening up wider applications for future works.
☆ Unified HT-CNNs Architecture: Transfer Learning for Segmenting Diverse Brain Tumors in MRI from Gliomas to Pediatric Tumors
Accurate segmentation of brain tumors from 3D multimodal MRI is vital for diagnosis and treatment planning across diverse brain tumors. This paper addresses the challenges posed by the BraTS 2023, presenting a unified transfer learning approach that applies to a broader spectrum of brain tumors. We introduce HT-CNNs, an ensemble of Hybrid Transformers and Convolutional Neural Networks optimized through transfer learning for varied brain tumor segmentation. This method captures spatial and contextual details from MRI data, fine-tuned on diverse datasets representing common tumor types. Through transfer learning, HT-CNNs utilize the learned representations from one task to improve generalization in another, harnessing the power of pre-trained models on large datasets and fine-tuning them on specific tumor types. We preprocess diverse datasets from multiple international distributions, ensuring representativeness for the most common brain tumors. Our rigorous evaluation employs standardized quantitative metrics across all tumor types, ensuring robustness and generalizability. The proposed ensemble model achieves superior segmentation results across the BraTS validation datasets over the previous winning methods. Comprehensive quantitative evaluations using the DSC and HD95 demonstrate the effectiveness of our approach. Qualitative segmentation predictions further validate the high-quality outputs produced by our model. Our findings underscore the potential of transfer learning and ensemble approaches in medical image segmentation, indicating a substantial enhancement in clinical decision-making and patient care. Despite facing challenges related to post-processing and domain gaps, our study sets a new precedent for future research for brain tumor segmentation. The docker image for the code and models has been made publicly available, https://hub.docker.com/r/razeineldin/ht-cnns.
comment: Accepted in the Computer Assisted Radiology and Surgery (CARS 2024) Conference
☆ Hierarchical Classification for Automated Image Annotation of Coral Reef Benthic Structures NeurIPS 2024
Automated benthic image annotation is crucial to efficiently monitor and protect coral reefs against climate change. Current machine learning approaches fail to capture the hierarchical nature of benthic organisms covering reef substrata, i.e., coral taxonomic levels and health condition. To address this limitation, we propose to annotate benthic images using hierarchical classification. Experiments on a custom dataset from a Northeast Brazilian coral reef show that our approach outperforms flat classifiers, improving both F1 and hierarchical F1 scores by approximately 2\% across varying amounts of training data. In addition, this hierarchical method aligns more closely with ecological objectives.
comment: Poster at Tackling Climate Change with Machine Learning: workshop at NeurIPS 2024
☆ Improving Active Learning with a Bayesian Representation of Epistemic Uncertainty
A popular strategy for active learning is to specifically target a reduction in epistemic uncertainty, since aleatoric uncertainty is often considered as being intrinsic to the system of interest and therefore not reducible. Yet, distinguishing these two types of uncertainty remains challenging and there is no single strategy that consistently outperforms the others. We propose to use a particular combination of probability and possibility theories, with the aim of using the latter to specifically represent epistemic uncertainty, and we show how this combination leads to new active learning strategies that have desirable properties. In order to demonstrate the efficiency of these strategies in non-trivial settings, we introduce the notion of a possibilistic Gaussian process (GP) and consider GP-based multiclass and binary classification problems, for which the proposed methods display a strong performance for both simulated and real datasets.
☆ Structured IB: Improving Information Bottleneck with Structured Feature Learning
The Information Bottleneck (IB) principle has emerged as a promising approach for enhancing the generalization, robustness, and interpretability of deep neural networks, demonstrating efficacy across image segmentation, document clustering, and semantic communication. Among IB implementations, the IB Lagrangian method, employing Lagrangian multipliers, is widely adopted. While numerous methods for the optimizations of IB Lagrangian based on variational bounds and neural estimators are feasible, their performance is highly dependent on the quality of their design, which is inherently prone to errors. To address this limitation, we introduce Structured IB, a framework for investigating potential structured features. By incorporating auxiliary encoders to extract missing informative features, we generate more informative representations. Our experiments demonstrate superior prediction accuracy and task-relevant information preservation compared to the original IB Lagrangian method, even with reduced network size.
☆ Generate Any Scene: Evaluating and Improving Text-to-Vision Generation with Scene Graph Programming
DALL-E and Sora have gained attention by producing implausible images, such as "astronauts riding a horse in space." Despite the proliferation of text-to-vision models that have inundated the internet with synthetic visuals, from images to 3D assets, current benchmarks predominantly evaluate these models on real-world scenes paired with captions. We introduce Generate Any Scene, a framework that systematically enumerates scene graphs representing a vast array of visual scenes, spanning realistic to imaginative compositions. Generate Any Scene leverages 'scene graph programming', a method for dynamically constructing scene graphs of varying complexity from a structured taxonomy of visual elements. This taxonomy includes numerous objects, attributes, and relations, enabling the synthesis of an almost infinite variety of scene graphs. Using these structured representations, Generate Any Scene translates each scene graph into a caption, enabling scalable evaluation of text-to-vision models through standard metrics. We conduct extensive evaluations across multiple text-to-image, text-to-video, and text-to-3D models, presenting key findings on model performance. We find that DiT-backbone text-to-image models align more closely with input captions than UNet-backbone models. Text-to-video models struggle with balancing dynamics and consistency, while both text-to-video and text-to-3D models show notable gaps in human preference alignment. We demonstrate the effectiveness of Generate Any Scene by conducting three practical applications leveraging captions generated by Generate Any Scene: 1) a self-improving framework where models iteratively enhance their performance using generated data, 2) a distillation process to transfer specific strengths from proprietary models to open-source counterparts, and 3) improvements in content moderation by identifying and generating challenging synthetic data.
☆ Model-Editing-Based Jailbreak against Safety-aligned Large Language Models
Large Language Models (LLMs) have transformed numerous fields by enabling advanced natural language interactions but remain susceptible to critical vulnerabilities, particularly jailbreak attacks. Current jailbreak techniques, while effective, often depend on input modifications, making them detectable and limiting their stealth and scalability. This paper presents Targeted Model Editing (TME), a novel white-box approach that bypasses safety filters by minimally altering internal model structures while preserving the model's intended functionalities. TME identifies and removes safety-critical transformations (SCTs) embedded in model matrices, enabling malicious queries to bypass restrictions without input modifications. By analyzing distinct activation patterns between safe and unsafe queries, TME isolates and approximates SCTs through an optimization process. Implemented in the D-LLM framework, our method achieves an average Attack Success Rate (ASR) of 84.86% on four mainstream open-source LLMs, maintaining high performance. Unlike existing methods, D-LLM eliminates the need for specific triggers or harmful response collections, offering a stealthier and more effective jailbreak strategy. This work reveals a covert and robust threat vector in LLM security and emphasizes the need for stronger safeguards in model safety alignment.
☆ Adaptive$^2$: Adaptive Domain Mining for Fine-grained Domain Adaptation Modeling
Advertising systems often face the multi-domain challenge, where data distributions vary significantly across scenarios. Existing domain adaptation methods primarily focus on building domain-adaptive neural networks but often rely on hand-crafted domain information, e.g., advertising placement, which may be sub-optimal. We think that fine-grained "domain" patterns exist that are difficult to hand-craft in online advertisement. Thus, we propose Adaptive$^2$, a novel framework that first learns domains adaptively using a domain mining module by self-supervision and then employs a shared&specific network to model shared and conflicting information. As a practice, we use VQ-VAE as the domain mining module and conduct extensive experiments on public benchmarks. Results show that traditional domain adaptation methods with hand-crafted domains perform no better than single-domain models under fair FLOPS conditions, highlighting the importance of domain definition. In contrast, Adaptive$^2$ outperforms existing approaches, emphasizing the effectiveness of our method and the significance of domain mining. We also deployed Adaptive$^2$ in the live streaming scenario of Kuaishou Advertising System, demonstrating its commercial value and potential for automatic domain identification. To the best of our knowledge, Adaptive$^2$ is the first approach to automatically learn both domain identification and adaptation in online advertising, opening new research directions for this area.
comment: 10 pages, 6 figures
☆ Magneto: Combining Small and Large Language Models for Schema Matching
Recent advances in language models opened new opportunities to address complex schema matching tasks. Schema matching approaches have been proposed that demonstrate the usefulness of language models, but they have also uncovered important limitations: Small language models (SLMs) require training data (which can be both expensive and challenging to obtain), and large language models (LLMs) often incur high computational costs and must deal with constraints imposed by context windows. We present Magneto, a cost-effective and accurate solution for schema matching that combines the advantages of SLMs and LLMs to address their limitations. By structuring the schema matching pipeline in two phases, retrieval and reranking, Magneto can use computationally efficient SLM-based strategies to derive candidate matches which can then be reranked by LLMs, thus making it possible to reduce runtime without compromising matching accuracy. We propose a self-supervised approach to fine-tune SLMs which uses LLMs to generate syntactically diverse training data, and prompting strategies that are effective for reranking. We also introduce a new benchmark, developed in collaboration with domain experts, which includes real biomedical datasets and presents new challenges to schema matching methods. Through a detailed experimental evaluation, using both our new and existing benchmarks, we show that Magneto is scalable and attains high accuracy for datasets from different domains.
☆ Mixture of Experts Meets Decoupled Message Passing: Towards General and Adaptive Node Classification
Graph neural networks excel at graph representation learning but struggle with heterophilous data and long-range dependencies. And graph transformers address these issues through self-attention, yet face scalability and noise challenges on large-scale graphs. To overcome these limitations, we propose GNNMoE, a universal model architecture for node classification. This architecture flexibly combines fine-grained message-passing operations with a mixture-of-experts mechanism to build feature encoding blocks. Furthermore, by incorporating soft and hard gating layers to assign the most suitable expert networks to each node, we enhance the model's expressive power and adaptability to different graph types. In addition, we introduce adaptive residual connections and an enhanced FFN module into GNNMoE, further improving the expressiveness of node representation. Extensive experimental results demonstrate that GNNMoE performs exceptionally well across various types of graph data, effectively alleviating the over-smoothing issue and global noise, enhancing model robustness and adaptability, while also ensuring computational efficiency on large-scale graphs.
comment: arXiv admin note: text overlap with arXiv:2410.11189
☆ From communities to interpretable network and word embedding: an unified approach
Modelling information from complex systems such as humans social interaction or words co-occurrences in our languages can help to understand how these systems are organized and function. Such systems can be modelled by networks, and network theory provides a useful set of methods to analyze them. Among these methods, graph embedding is a powerful tool to summarize the interactions and topology of a network in a vectorized feature space. When used in input of machine learning algorithms, embedding vectors help with common graph problems such as link prediction, graph matching, etc. Word embedding has the goal of representing the sense of words, extracting it from large text corpora. Despite differences in the structure of information in input of embedding algorithms, many graph embedding approaches are adapted and inspired from methods in NLP. Limits of these methods are observed in both domains. Most of these methods require long and resource greedy training. Another downside to most methods is that they are black-box, from which understanding how the information is structured is rather complex. Interpretability of a model allows understanding how the vector space is structured without the need for external information, and thus can be audited more easily. With both these limitations in mind, we propose a novel framework to efficiently embed network vertices in an interpretable vector space. Our Lower Dimension Bipartite Framework (LDBGF) leverages the bipartite projection of a network using cliques to reduce dimensionality. Along with LDBGF, we introduce two implementations of this framework that rely on communities instead of cliques: SINr-NR and SINr-MF. We show that SINr-MF can perform well on classical graphs and SINr-NR can produce high-quality graph and word embeddings that are interpretable and stable across runs.
☆ Analyzing and Improving Model Collapse in Rectified Flow Models
Generative models aim to produce synthetic data indistinguishable from real distributions, but iterative training on self-generated data can lead to \emph{model collapse (MC)}, where performance degrades over time. In this work, we provide the first theoretical analysis of MC in Rectified Flow by framing it within the context of Denoising Autoencoders (DAEs). We show that when DAE models are trained on recursively generated synthetic data with small noise variance, they suffer from MC with progressive diminishing generation quality. To address this MC issue, we propose methods that strategically incorporate real data into the training process, even when direct noise-image pairs are unavailable. Our proposed techniques, including Reverse Collapse-Avoiding (RCA) Reflow and Online Collapse-Avoiding Reflow (OCAR), effectively prevent MC while maintaining the efficiency benefits of Rectified Flow. Extensive experiments on standard image datasets demonstrate that our methods not only mitigate MC but also improve sampling efficiency, leading to higher-quality image generation with fewer sampling steps.
☆ Can Graph Neural Networks Learn Language with Extremely Weak Text Supervision?
While great success has been achieved in building vision models with Contrastive Language-Image Pre-training (CLIP) over Internet-scale image-text pairs, building transferable Graph Neural Networks (GNNs) with CLIP pipeline is challenging because of three fundamental issues: the scarcity of labeled data and text supervision, different levels of downstream tasks, and the conceptual gaps between domains. In this work, to address these issues, we leverage multi-modal prompt learning to effectively adapt pre-trained GNN to downstream tasks and data, given only a few semantically labeled samples, each with extremely weak text supervision. Our new paradigm embeds the graphs directly in the same space as the Large Language Models (LLMs) by learning both graph prompts and text prompts simultaneously. To accomplish this, we improve state-of-the-art graph prompt method, and then propose the first graph-language multi-modal prompt learning approach for exploiting the knowledge in pre-trained models. Notably, due to the insufficient supervision for fine-tuning, in our paradigm, the pre-trained GNN and the LLM are kept frozen, so the learnable parameters are much fewer than fine-tuning any pre-trained model. Through extensive experiments on real-world datasets, we demonstrate the superior performance of our paradigm in few-shot, multi-task-level, and cross-domain settings. Moreover, we build the first CLIP-style zero-shot classification prototype that can generalize GNNs to unseen classes with extremely weak text supervision.
comment: Preprint, 26 pages
☆ Diversity Drives Fairness: Ensemble of Higher Order Mutants for Intersectional Fairness of Machine Learning Software ICSE 2025
Intersectional fairness is a critical requirement for Machine Learning (ML) software, demanding fairness across subgroups defined by multiple protected attributes. This paper introduces FairHOME, a novel ensemble approach using higher order mutation of inputs to enhance intersectional fairness of ML software during the inference phase. Inspired by social science theories highlighting the benefits of diversity, FairHOME generates mutants representing diverse subgroups for each input instance, thus broadening the array of perspectives to foster a fairer decision-making process. Unlike conventional ensemble methods that combine predictions made by different models, FairHOME combines predictions for the original input and its mutants, all generated by the same ML model, to reach a final decision. Notably, FairHOME is even applicable to deployed ML software as it bypasses the need for training new models. We extensively evaluate FairHOME against seven state-of-the-art fairness improvement methods across 24 decision-making tasks using widely adopted metrics. FairHOME consistently outperforms existing methods across all metrics considered. On average, it enhances intersectional fairness by 47.5%, surpassing the currently best-performing method by 9.6 percentage points.
comment: Accepted by the 47th International Conference on Software Engineering (ICSE 2025). Please include ICSE in any citations
☆ Collaborative Hybrid Propagator for Temporal Misalignment in Audio-Visual Segmentation
Audio-visual video segmentation (AVVS) aims to generate pixel-level maps of sound-producing objects that accurately align with the corresponding audio. However, existing methods often face temporal misalignment, where audio cues and segmentation results are not temporally coordinated. Audio provides two critical pieces of information: i) target object-level details and ii) the timing of when objects start and stop producing sounds. Current methods focus more on object-level information but neglect the boundaries of audio semantic changes, leading to temporal misalignment. To address this issue, we propose a Collaborative Hybrid Propagator Framework~(Co-Prop). This framework includes two main steps: Preliminary Audio Boundary Anchoring and Frame-by-Frame Audio-Insert Propagation. To Anchor the audio boundary, we employ retrieval-assist prompts with Qwen large language models to identify control points of audio semantic changes. These control points split the audio into semantically consistent audio portions. After obtaining the control point lists, we propose the Audio Insertion Propagator to process each audio portion using a frame-by-frame audio insertion propagation and matching approach. We curated a compact dataset comprising diverse source conversion cases and devised a metric to assess alignment rates. Compared to traditional simultaneous processing methods, our approach reduces memory requirements and facilitates frame alignment. Experimental results demonstrate the effectiveness of our approach across three datasets and two backbones. Furthermore, our method can be integrated with existing AVVS approaches, offering plug-and-play functionality to enhance their performance.
☆ DG-Mamba: Robust and Efficient Dynamic Graph Structure Learning with Selective State Space Models AAAI
Dynamic graphs exhibit intertwined spatio-temporal evolutionary patterns, widely existing in the real world. Nevertheless, the structure incompleteness, noise, and redundancy result in poor robustness for Dynamic Graph Neural Networks (DGNNs). Dynamic Graph Structure Learning (DGSL) offers a promising way to optimize graph structures. However, aside from encountering unacceptable quadratic complexity, it overly relies on heuristic priors, making it hard to discover underlying predictive patterns. How to efficiently refine the dynamic structures, capture intrinsic dependencies, and learn robust representations, remains under-explored. In this work, we propose the novel DG-Mamba, a robust and efficient Dynamic Graph structure learning framework with the Selective State Space Models (Mamba). To accelerate the spatio-temporal structure learning, we propose a kernelized dynamic message-passing operator that reduces the quadratic time complexity to linear. To capture global intrinsic dynamics, we establish the dynamic graph as a self-contained system with State Space Model. By discretizing the system states with the cross-snapshot graph adjacency, we enable the long-distance dependencies capturing with the selective snapshot scan. To endow learned dynamic structures more expressive with informativeness, we propose the self-supervised Principle of Relevant Information for DGSL to regularize the most relevant yet least redundant information, enhancing global robustness. Extensive experiments demonstrate the superiority of the robustness and efficiency of our DG-Mamba compared with the state-of-the-art baselines against adversarial attacks.
comment: Accepted by the Main Technical Track of the 39th Annual AAAI Conference on Artificial Intelligence (AAAI-2025)
☆ How Vision-Language Tasks Benefit from Large Pre-trained Models: A Survey
The exploration of various vision-language tasks, such as visual captioning, visual question answering, and visual commonsense reasoning, is an important area in artificial intelligence and continuously attracts the research community's attention. Despite the improvements in overall performance, classic challenges still exist in vision-language tasks and hinder the development of this area. In recent years, the rise of pre-trained models is driving the research on vision-language tasks. Thanks to the massive scale of training data and model parameters, pre-trained models have exhibited excellent performance in numerous downstream tasks. Inspired by the powerful capabilities of pre-trained models, new paradigms have emerged to solve the classic challenges. Such methods have become mainstream in current research with increasing attention and rapid advances. In this paper, we present a comprehensive overview of how vision-language tasks benefit from pre-trained models. First, we review several main challenges in vision-language tasks and discuss the limitations of previous solutions before the era of pre-training. Next, we summarize the recent advances in incorporating pre-trained models to address the challenges in vision-language tasks. Finally, we analyze the potential risks associated with the inherent limitations of pre-trained models and discuss possible solutions, attempting to provide future research directions.
comment: Under Review
☆ How to Weight Multitask Finetuning? Fast Previews via Bayesian Model-Merging
When finetuning multiple tasks altogether, it is important to carefully weigh them to get a good performance, but searching for good weights can be difficult and costly. Here, we propose to aid the search with fast previews to quickly get a rough idea of different reweighting options. We use model merging to create previews by simply reusing and averaging parameters of models trained on each task separately (no retraining required). To improve the quality of previews, we propose a Bayesian approach to design new merging strategies by using more flexible posteriors. We validate our findings on vision and natural-language transformers. Our work shows the benefits of model merging via Bayes to improve multitask finetuning.
☆ AGMixup: Adaptive Graph Mixup for Semi-supervised Node Classification AAAI 2025
Mixup is a data augmentation technique that enhances model generalization by interpolating between data points using a mixing ratio $\lambda$ in the image domain. Recently, the concept of mixup has been adapted to the graph domain through node-centric interpolations. However, these approaches often fail to address the complexity of interconnected relationships, potentially damaging the graph's natural topology and undermining node interactions. Furthermore, current graph mixup methods employ a one-size-fits-all strategy with a randomly sampled $\lambda$ for all mixup pairs, ignoring the diverse needs of different pairs. This paper proposes an Adaptive Graph Mixup (AGMixup) framework for semi-supervised node classification. AGMixup introduces a subgraph-centric approach, which treats each subgraph similarly to how images are handled in Euclidean domains, thus facilitating a more natural integration of mixup into graph-based learning. We also propose an adaptive mechanism to tune the mixing ratio $\lambda$ for diverse mixup pairs, guided by the contextual similarity and uncertainty of the involved subgraphs. Extensive experiments across seven datasets on semi-supervised node classification benchmarks demonstrate AGMixup's superiority over state-of-the-art graph mixup methods. Source codes are available at \url{https://github.com/WeigangLu/AGMixup}.
comment: Accepted by AAAI 2025
☆ Wasserstein Distance Rivals Kullback-Leibler Divergence for Knowledge Distillation NeurIPS 2024
Since pioneering work of Hinton et al., knowledge distillation based on Kullback-Leibler Divergence (KL-Div) has been predominant, and recently its variants have achieved compelling performance. However, KL-Div only compares probabilities of the corresponding category between the teacher and student while lacking a mechanism for cross-category comparison. Besides, KL-Div is problematic when applied to intermediate layers, as it cannot handle non-overlapping distributions and is unaware of geometry of the underlying manifold. To address these downsides, we propose a methodology of Wasserstein Distance (WD) based knowledge distillation. Specifically, we propose a logit distillation method called WKD-L based on discrete WD, which performs cross-category comparison of probabilities and thus can explicitly leverage rich interrelations among categories. Moreover, we introduce a feature distillation method called WKD-F, which uses a parametric method for modeling feature distributions and adopts continuous WD for transferring knowledge from intermediate layers. Comprehensive evaluations on image classification and object detection have shown (1) for logit distillation WKD-L outperforms very strong KL-Div variants; (2) for feature distillation WKD-F is superior to the KL-Div counterparts and state-of-the-art competitors. The source code is available at https://peihuali.org/WKD
comment: Accepted to NeurIPS 2024. Equal contribution from first two authors
☆ Learn How to Query from Unlabeled Data Streams in Federated Learning
Federated learning (FL) enables collaborative learning among decentralized clients while safeguarding the privacy of their local data. Existing studies on FL typically assume offline labeled data available at each client when the training starts. Nevertheless, the training data in practice often arrive at clients in a streaming fashion without ground-truth labels. Given the expensive annotation cost, it is critical to identify a subset of informative samples for labeling on clients. However, selecting samples locally while accommodating the global training objective presents a challenge unique to FL. In this work, we tackle this conundrum by framing the data querying process in FL as a collaborative decentralized decision-making problem and proposing an effective solution named LeaDQ, which leverages multi-agent reinforcement learning algorithms. In particular, under the implicit guidance from global information, LeaDQ effectively learns the local policies for distributed clients and steers them towards selecting samples that can enhance the global model's accuracy. Extensive simulations on image and text tasks show that LeaDQ advances the model performance in various FL scenarios, outperforming the benchmarking algorithms.
☆ DOGE: An Extrinsic Orientation and Gyroscope Bias Estimation for Visual-Inertial Odometry Initialization
Most existing visual-inertial odometry (VIO) initialization methods rely on accurate pre-calibrated extrinsic parameters. However, during long-term use, irreversible structural deformation caused by temperature changes, mechanical squeezing, etc. will cause changes in extrinsic parameters, especially in the rotational part. Existing initialization methods that simultaneously estimate extrinsic parameters suffer from poor robustness, low precision, and long initialization latency due to the need for sufficient translational motion. To address these problems, we propose a novel VIO initialization method, which jointly considers extrinsic orientation and gyroscope bias within the normal epipolar constraints, achieving higher precision and better robustness without delayed rotational calibration. First, a rotation-only constraint is designed for extrinsic orientation and gyroscope bias estimation, which tightly couples gyroscope measurements and visual observations and can be solved in pure-rotation cases. Second, we propose a weighting strategy together with a failure detection strategy to enhance the precision and robustness of the estimator. Finally, we leverage Maximum A Posteriori to refine the results before enough translation parallax comes. Extensive experiments have demonstrated that our method outperforms the state-of-the-art methods in both accuracy and robustness while maintaining competitive efficiency.
☆ DiffRaman: A Conditional Latent Denoising Diffusion Probabilistic Model for Bacterial Raman Spectroscopy Identification Under Limited Data Conditions
Raman spectroscopy has attracted significant attention in various biochemical detection fields, especially in the rapid identification of pathogenic bacteria. The integration of this technology with deep learning to facilitate automated bacterial Raman spectroscopy diagnosis has emerged as a key focus in recent research. However, the diagnostic performance of existing deep learning methods largely depends on a sufficient dataset, and in scenarios where there is a limited availability of Raman spectroscopy data, it is inadequate to fully optimize the numerous parameters of deep neural networks. To address these challenges, this paper proposes a data generation method utilizing deep generative models to expand the data volume and enhance the recognition accuracy of bacterial Raman spectra. Specifically, we introduce DiffRaman, a conditional latent denoising diffusion probability model for Raman spectra generation. Experimental results demonstrate that synthetic bacterial Raman spectra generated by DiffRaman can effectively emulate real experimental spectra, thereby enhancing the performance of diagnostic models, especially under conditions of limited data. Furthermore, compared to existing generative models, the proposed DiffRaman offers improvements in both generation quality and computational efficiency. Our DiffRaman approach offers a well-suited solution for automated bacteria Raman spectroscopy diagnosis in data-scarce scenarios, offering new insights into alleviating the labor of spectroscopic measurements and enhancing rare bacteria identification.
☆ Why Does Dropping Edges Usually Outperform Adding Edges in Graph Contrastive Learning?
Graph contrastive learning (GCL) has been widely used as an effective self-supervised learning method for graph representation learning. However, how to apply adequate and stable graph augmentation to generating proper views for contrastive learning remains an essential problem. Dropping edges is a primary augmentation in GCL while adding edges is not a common method due to its unstable performance. To our best knowledge, there is no theoretical analysis to study why dropping edges usually outperforms adding edges. To answer this question, we introduce a new metric, namely Error Passing Rate (EPR), to quantify how a graph fits the network. Inspired by the theoretical conclusions, we propose a novel GCL algorithm, Error-PAssing-based Graph Contrastive Learning (EPAGCL), which uses both edge adding and edge dropping as its augmentation. To be specific, we generate views by adding and dropping edges according to the weights derived from EPR. Extensive experiments on various real-world datasets are conducted to validate the correctness of our theoretical analysis and the effectiveness of our proposed algorithm.
☆ Evil twins are not that evil: Qualitative insights into machine-generated prompts
It has been widely observed that language models (LMs) respond in predictable ways to algorithmically generated prompts that are seemingly unintelligible. This is both a sign that we lack a full understanding of how LMs work, and a practical challenge, because opaqueness can be exploited for harmful uses of LMs, such as jailbreaking. We present the first thorough analysis of opaque machine-generated prompts, or autoprompts, pertaining to 3 LMs of different sizes and families. We find that machine-generated prompts are characterized by a last token that is often intelligible and strongly affects the generation. A small but consistent proportion of the previous tokens are fillers that probably appear in the prompt as a by-product of the fact that the optimization process fixes the number of tokens. The remaining tokens tend to have at least a loose semantic relation with the generation, although they do not engage in well-formed syntactic relations with it. We find moreover that some of the ablations we applied to machine-generated prompts can also be applied to natural language sequences, leading to similar behavior, suggesting that autoprompts are a direct consequence of the way in which LMs process linguistic inputs in general.
☆ Progressive Multi-granular Alignments for Grounded Reasoning in Large Vision-Language Models
Existing Large Vision-Language Models (LVLMs) excel at matching concepts across multi-modal inputs but struggle with compositional concepts and high-level relationships between entities. This paper introduces Progressive multi-granular Vision-Language alignments (PromViL), a novel framework to enhance LVLMs' ability in performing grounded compositional visual reasoning tasks. Our approach constructs a hierarchical structure of multi-modal alignments, ranging from simple to complex concepts. By progressively aligning textual descriptions with corresponding visual regions, our model learns to leverage contextual information from lower levels to inform higher-level reasoning. To facilitate this learning process, we introduce a data generation process that creates a novel dataset derived from Visual Genome, providing a wide range of nested compositional vision-language pairs. Experimental results demonstrate that our PromViL framework significantly outperforms baselines on various visual grounding and compositional question answering tasks.
☆ Dense Depth from Event Focal Stack WACV2025
We propose a method for dense depth estimation from an event stream generated when sweeping the focal plane of the driving lens attached to an event camera. In this method, a depth map is inferred from an ``event focal stack'' composed of the event stream using a convolutional neural network trained with synthesized event focal stacks. The synthesized event stream is created from a focal stack generated by Blender for any arbitrary 3D scene. This allows for training on scenes with diverse structures. Additionally, we explored methods to eliminate the domain gap between real event streams and synthetic event streams. Our method demonstrates superior performance over a depth-from-defocus method in the image domain on synthetic and real datasets.
comment: Accepted at WACV2025
♻ ☆ Reducing Reasoning Costs -- The Path of Optimization for Chain of Thought via Sparse Attention Mechanism NeurIPS 2024
In order to address the chain of thought in the large language model inference cost surge, this research proposes to use a sparse attention mechanism that only focuses on a few relevant tokens. The researcher constructed a new attention mechanism and used GiantRabbit trained with custom GPTs as an experimental tool. The experiment tested and compared the reasoning time, correctness score and chain of thought length of this model and o1 Preview in solving the linear algebra test questions of MIT OpenCourseWare. The results show that GiantRabbit's reasoning time and chain of thought length are significantly lower than o1 Preview. It verifies the feasibility of sparse attention mechanism for optimizing chain of thought reasoning. Detailed architectural details and experimental process have been uploaded to Github, the link is:https://github.com/brucewang123456789/GeniusTrail.git.
comment: The main text is 5 pages, totaling 9 pages; 4 figures, 1 table. It have been submitted to NeurIPS 2024 Workshop MusIML and OpenReview
♻ ☆ Length Optimization in Conformal Prediction
Conditional validity and length efficiency are two crucial aspects of conformal prediction (CP). Conditional validity ensures accurate uncertainty quantification for data subpopulations, while proper length efficiency ensures that the prediction sets remain informative. Despite significant efforts to address each of these issues individually, a principled framework that reconciles these two objectives has been missing in the CP literature. In this paper, we develop Conformal Prediction with Length-Optimization (CPL) - a novel and practical framework that constructs prediction sets with (near-) optimal length while ensuring conditional validity under various classes of covariate shifts, including the key cases of marginal and group-conditional coverage. In the infinite sample regime, we provide strong duality results which indicate that CPL achieves conditional validity and length optimality. In the finite sample regime, we show that CPL constructs conditionally valid prediction sets. Our extensive empirical evaluations demonstrate the superior prediction set size performance of CPL compared to state-of-the-art methods across diverse real-world and synthetic datasets in classification, regression, and large language model-based multiple choice question answering. An Implementation of our algorithm can be accessed at the following link: https://github.com/shayankiyani98/CP.
♻ ☆ Concept Bottleneck Language Models For protein design
We introduce Concept Bottleneck Protein Language Models (CB-pLM), a generative masked language model with a layer where each neuron corresponds to an interpretable concept. Our architecture offers three key benefits: i) Control: We can intervene on concept values to precisely control the properties of generated proteins, achieving a 3 times larger change in desired concept values compared to baselines. ii) Interpretability: A linear mapping between concept values and predicted tokens allows transparent analysis of the model's decision-making process. iii) Debugging: This transparency facilitates easy debugging of trained models. Our models achieve pre-training perplexity and downstream task performance comparable to traditional masked protein language models, demonstrating that interpretability does not compromise performance. While adaptable to any language model, we focus on masked protein language models due to their importance in drug discovery and the ability to validate our model's capabilities through real-world experiments and expert knowledge. We scale our CB-pLM from 24 million to 3 billion parameters, making them the largest Concept Bottleneck Models trained and the first capable of generative language modeling.
♻ ☆ Efficient Online Reinforcement Learning Fine-Tuning Need Not Retain Offline Data
The modern paradigm in machine learning involves pre-training on diverse data, followed by task-specific fine-tuning. In reinforcement learning (RL), this translates to learning via offline RL on a diverse historical dataset, followed by rapid online RL fine-tuning using interaction data. Most RL fine-tuning methods require continued training on offline data for stability and performance. However, this is undesirable because training on diverse offline data is slow and expensive for large datasets, and in principle, also limit the performance improvement possible because of constraints or pessimism on offline data. In this paper, we show that retaining offline data is unnecessary as long as we use a properly-designed online RL approach for fine-tuning offline RL initializations. To build this approach, we start by analyzing the role of retaining offline data in online fine-tuning. We find that continued training on offline data is mostly useful for preventing a sudden divergence in the value function at the onset of fine-tuning, caused by a distribution mismatch between the offline data and online rollouts. This divergence typically results in unlearning and forgetting the benefits of offline pre-training. Our approach, Warm-start RL (WSRL), mitigates the catastrophic forgetting of pre-trained initializations using a very simple idea. WSRL employs a warmup phase that seeds the online RL run with a very small number of rollouts from the pre-trained policy to do fast online RL. The data collected during warmup helps ``recalibrate'' the offline Q-function to the online distribution, allowing us to completely discard offline data without destabilizing the online RL fine-tuning. We show that WSRL is able to fine-tune without retaining any offline data, and is able to learn faster and attains higher performance than existing algorithms irrespective of whether they retain offline data or not.
♻ ☆ Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Change KDD
Counterfactual explanations (CFEs) guide users on how to adjust inputs to machine learning models to achieve desired outputs. While existing research primarily addresses static scenarios, real-world applications often involve data or model changes, potentially invalidating previously generated CFEs and rendering user-induced input changes ineffective. Current methods addressing this issue often support only specific models or change types, require extensive hyperparameter tuning, or fail to provide probabilistic guarantees on CFE robustness to model changes. This paper proposes a novel approach for generating CFEs that provides probabilistic guarantees for any model and change type, while offering interpretable and easy-to-select hyperparameters. We establish a theoretical framework for probabilistically defining robustness to model change and demonstrate how our BetaRCE method directly stems from it. BetaRCE is a post-hoc method applied alongside a chosen base CFE generation method to enhance the quality of the explanation beyond robustness. It facilitates a transition from the base explanation to a more robust one with user-adjusted probability bounds. Through experimental comparisons with baselines, we show that BetaRCE yields robust, most plausible, and closest to baseline counterfactual explanations.
comment: Accepted at 31st SIGKDD Conference on Knowledge Discovery and Data Mining 2025
♻ ☆ Fusing Domain-Specific Content from Large Language Models into Knowledge Graphs for Enhanced Zero Shot Object State Classification AAAI
Domain-specific knowledge can significantly contribute to addressing a wide variety of vision tasks. However, the generation of such knowledge entails considerable human labor and time costs. This study investigates the potential of Large Language Models (LLMs) in generating and providing domain-specific information through semantic embeddings. To achieve this, an LLM is integrated into a pipeline that utilizes Knowledge Graphs and pre-trained semantic vectors in the context of the Vision-based Zero-shot Object State Classification task. We thoroughly examine the behavior of the LLM through an extensive ablation study. Our findings reveal that the integration of LLM-based embeddings, in combination with general-purpose pre-trained embeddings, leads to substantial performance improvements. Drawing insights from this ablation study, we conduct a comparative analysis against competing models, thereby highlighting the state-of-the-art performance achieved by the proposed approach.
comment: Accepted at the AAAI-MAKE 2024
♻ ☆ Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Prompts
As large language models (LLMs) become increasingly prevalent across many real-world applications, understanding and enhancing their robustness to adversarial attacks is of paramount importance. Existing methods for identifying adversarial prompts tend to focus on specific domains, lack diversity, or require extensive human annotations. To address these limitations, we present Rainbow Teaming, a novel black-box approach for producing a diverse collection of adversarial prompts. Rainbow Teaming casts adversarial prompt generation as a quality-diversity problem and uses open-ended search to generate prompts that are both effective and diverse. Focusing on the safety domain, we use Rainbow Teaming to target various state-of-the-art LLMs, including the Llama 2 and Llama 3 models. Our approach reveals hundreds of effective adversarial prompts, with an attack success rate exceeding 90% across all tested models. Furthermore, we demonstrate that prompts generated by Rainbow Teaming are highly transferable and that fine-tuning models with synthetic data generated by our method significantly enhances their safety without sacrificing general performance or helpfulness. We additionally explore the versatility of Rainbow Teaming by applying it to question answering and cybersecurity, showcasing its potential to drive robust open-ended self-improvement in a wide range of applications.
♻ ☆ Harnessing the Power of Vicinity-Informed Analysis for Classification under Covariate Shift
Transfer learning enhances prediction accuracy on a target distribution by leveraging data from a source distribution, demonstrating significant benefits in various applications. This paper introduces a novel dissimilarity measure that utilizes vicinity information, i.e., the local structure of data points, to analyze the excess error in classification under covariate shift, a transfer learning setting where marginal feature distributions differ but conditional label distributions remain the same. We characterize the excess error using the proposed measure and demonstrate faster or competitive convergence rates compared to previous techniques. Notably, our approach is effective in the support non-containment assumption, which often appears in real-world applications, holds. Our theoretical analysis bridges the gap between current theoretical findings and empirical observations in transfer learning, particularly in scenarios with significant differences between source and target distributions.
♻ ☆ Continuous Multidimensional Scaling
Multidimensional scaling (MDS) is the act of embedding proximity information about a set of $n$ objects in $d$-dimensional Euclidean space. As originally conceived by the psychometric community, MDS was concerned with embedding a fixed set of proximities associated with a fixed set of objects. Modern concerns, e.g., that arise in developing asymptotic theories for statistical inference on random graphs, more typically involve studying the limiting behavior of a sequence of proximities associated with an increasing set of objects. Here we are concerned with embedding dissimilarities by minimizing Kruskal's (1964) raw stress criterion. Standard results from the theory of point-to-set maps can be used to establish that, if $n$ is fixed and a sequence of dissimilarity matrices converges, then the limit of their embedded structures is the embedded structure of the limiting dissimilarity matrix. But what if $n$ increases? It then becomes necessary to reformulate MDS so that the entire sequence of embedding problems can be viewed as a sequence of optimization problems in a fixed space. We present such a reformulation, {\em continuous MDS}. Within the continuous MDS framework, we derive two $L^p$ consistency results, one for embedding without constraints on the configuration, the other for embedding subject to {\em approximate Lipschitz constraints}\/ that encourage smoothness of the embedding function. The latter approach, {\em Approximate Lipschitz Embedding}\/ (ALE) is new. Finally, we demonstrate that embedded structures produced by ALE can be interpolated in a way that results in uniform convergence.
comment: 25 pages. Modified previous material for greater clarity; added new material about approximate Lipschitz constraints, Approximate Lipschitz Embedding (ALE), and uniform convergence; added material on constrained minimization of raw stress to the appendix
♻ ☆ Correct-N-Contrast: A Contrastive Approach for Improving Robustness to Spurious Correlations ICML 2022
Spurious correlations pose a major challenge for robust machine learning. Models trained with empirical risk minimization (ERM) may learn to rely on correlations between class labels and spurious attributes, leading to poor performance on data groups without these correlations. This is particularly challenging to address when spurious attribute labels are unavailable. To improve worst-group performance on spuriously correlated data without training attribute labels, we propose Correct-N-Contrast (CNC), a contrastive approach to directly learn representations robust to spurious correlations. As ERM models can be good spurious attribute predictors, CNC works by (1) using a trained ERM model's outputs to identify samples with the same class but dissimilar spurious features, and (2) training a robust model with contrastive learning to learn similar representations for same-class samples. To support CNC, we introduce new connections between worst-group error and a representation alignment loss that CNC aims to minimize. We empirically observe that worst-group error closely tracks with alignment loss, and prove that the alignment loss over a class helps upper-bound the class's worst-group vs. average error gap. On popular benchmarks, CNC reduces alignment loss drastically, and achieves state-of-the-art worst-group accuracy by 3.6% average absolute lift. CNC is also competitive with oracle methods that require group labels.
comment: 38 pages, 14 figures. ICML 2022 Long Talk
♻ ☆ Estimating the Number of HTTP/3 Responses in QUIC Using Deep Learning
QUIC, a new and increasingly used transport protocol, enhances TCP by offering improved security, performance, and stream multiplexing. These features, however, also impose challenges for network middle-boxes that need to monitor and analyze web traffic. This paper proposes a novel method to estimate the number of HTTP/3 responses in a given QUIC connection by an observer. This estimation reveals server behavior, client-server interactions, and data transmission efficiency, which is crucial for various applications such as designing a load balancing solution and detecting HTTP/3 flood attacks. The proposed scheme transforms QUIC connection traces into image sequences and uses machine learning (ML) models, guided by a tailored loss function, to predict response counts. Evaluations on more than seven million images-derived from 100,000 traces collected across 44,000 websites over four months-achieve up to 97% accuracy in both known and unknown server settings and 92% accuracy on previously unseen complete QUIC traces.
comment: arXiv admin note: substantial text overlap with arXiv:2410.03728
♻ ☆ Robust Multi-Agent Control via Maximum Entropy Heterogeneous-Agent Reinforcement Learning
In multi-agent reinforcement learning, optimal control with robustness guarantees are critical for its deployment in real world. However, existing methods face challenges related to sample complexity, training instability, potential suboptimal Nash Equilibrium convergence and non-robustness to multiple perturbations. In this paper, we propose a unified framework for learning \emph{stochastic} policies to resolve these issues. We embed cooperative MARL problems into probabilistic graphical models, from which we derive the maximum entropy (MaxEnt) objective optimal for MARL. Based on the MaxEnt framework, we propose \emph{Heterogeneous-Agent Soft Actor-Critic} (HASAC) algorithm. Theoretically, we prove the monotonic improvement and convergence to \emph{quantal response equilibrium} (QRE) properties of HASAC. Furthermore, HASAC is provably robust against a wide range of real-world uncertainties, including perturbations in rewards, environment dynamics, states, and actions. Finally, we generalize a unified template for MaxEnt algorithmic design named \emph{Maximum Entropy Heterogeneous-Agent Mirror Learning} (MEHAML), which provides any induced method with the same guarantees as HASAC. We evaluate HASAC on seven benchmarks: Bi-DexHands, Multi-Agent MuJoCo, Pursuit-Evade, StarCraft Multi-Agent Challenge, Google Research Football, Multi-Agent Particle Environment, Light Aircraft Game. Results show that HASAC consistently outperforms strong baselines in 34 out of 38 tasks, exhibiting improved training stability, better sample efficiency and sufficient exploration. The robustness of HASAC was further validated when encountering uncertainties in rewards, dynamics, states, and actions of 14 magnitudes, and real-world deployment in a multi-robot arena against these four types of uncertainties. See our page at \url{https://sites.google.com/view/meharl}.
comment: Work in Progress
♻ ☆ The BrowserGym Ecosystem for Web Agent Research
The BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents, particularly those leveraging automation and Large Language Models (LLMs) for web interaction tasks. Many existing benchmarks suffer from fragmentation and inconsistent evaluation methodologies, making it challenging to achieve reliable comparisons and reproducible results. BrowserGym aims to solve this by providing a unified, gym-like environment with well-defined observation and action spaces, facilitating standardized evaluation across diverse benchmarks. Combined with AgentLab, a complementary framework that aids in agent creation, testing, and analysis, BrowserGym offers flexibility for integrating new benchmarks while ensuring consistent evaluation and comprehensive experiment management. This standardized approach seeks to reduce the time and complexity of developing web agents, supporting more reliable comparisons and facilitating in-depth analysis of agent behaviors, and could result in more adaptable, capable agents, ultimately accelerating innovation in LLM-driven automation. As a supporting evidence, we conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across all benchmarks currently available in BrowserGym. Among other findings, our results highlight a large discrepancy between OpenAI and Anthropic's latests models, with Claude-3.5-Sonnet leading the way on almost all benchmarks, except on vision-related tasks where GPT-4o is superior. Despite these advancements, our results emphasize that building robust and efficient web agents remains a significant challenge, due to the inherent complexity of real-world web environments and the limitations of current models.
♻ ☆ Differentially Private Low-dimensional Synthetic Data from High-dimensional Datasets
Differentially private synthetic data provide a powerful mechanism to enable data analysis while protecting sensitive information about individuals. However, when the data lie in a high-dimensional space, the accuracy of the synthetic data suffers from the curse of dimensionality. In this paper, we propose a differentially private algorithm to generate low-dimensional synthetic data efficiently from a high-dimensional dataset with a utility guarantee with respect to the Wasserstein distance. A key step of our algorithm is a private principal component analysis (PCA) procedure with a near-optimal accuracy bound that circumvents the curse of dimensionality. Unlike the standard perturbation analysis, our analysis of private PCA works without assuming the spectral gap for the covariance matrix.
comment: 23 pages
♻ ☆ MathDSL: A Domain-Specific Language for Concise Mathematical Solutions Via Program Synthesis
We present MathDSL, a Domain-Specific Language (DSL) for mathematical equation solving, which, when deployed in program synthesis models, outperforms state-of-the-art reinforcement-learning-based methods. We also introduce a quantitative metric for measuring the conciseness of a mathematical solution and demonstrate the improvement in the quality of generated solutions compared to other methods. Our system demonstrates that a program synthesis system (DreamCoder) using MathDSL can generate programs that solve linear equations with greater accuracy and conciseness than using reinforcement learning systems. Additionally, we demonstrate that if we use the action spaces of previous reinforcement learning systems as DSLs, MathDSL outperforms the action-space-DSLs. We use DreamCoder to store equation-solving strategies as learned abstractions in its program library and demonstrate that by using MathDSL, these can be converted into human-interpretable solution strategies that could have applications in mathematical education.
comment: There was a typo in Figure 1 (a step in the Lemma solution was accidentally included twice). Additionally, our final experiment runs have MathDSL using one less step for this question, and ConPoLe using one more step to differentiate a division and a fraction in its final solution. Figure 1 has been updated to provide an exact copy of the experiment runs in the GitHub repository
♻ ☆ Self-Interested Agents in Collaborative Learning: An Incentivized Adaptive Data-Centric Framework
We propose a framework for adaptive data-centric collaborative learning among self-interested agents, coordinated by an arbiter. Designed to handle the incremental nature of real-world data, the framework operates in an online manner: at each step, the arbiter collects a batch of data from agents, trains a machine learning model, and provides each agent with a distinct model reflecting its data contributions. This setup establishes a feedback loop where shared data influence model updates, and the resulting models guide future data-sharing strategies. Agents evaluate and partition their data, selecting a partition to share using a stochastic parameterized policy optimized via policy gradient methods to optimize the utility of the received model as defined by agent-specific evaluation functions. On the arbiter side, the expected loss function over the true data distribution is optimized, incorporating agent-specific weights to account for distributional differences arising from diverse sources and selective sharing. A bilevel optimization algorithm jointly learns the model parameters and agent-specific weights. Mean-zero noise, computed using a distortion function that adjusts these agent-specific weights, is introduced to generate distinct agent-specific models, promoting valuable data sharing without requiring separate training. Our framework is underpinned by non-asymptotic analyses, ensuring convergence of the agent-side policy optimization to an approximate stationary point of the evaluation functions and convergence of the arbiter-side optimization to an approximate stationary point of the expected loss function.
♻ ☆ Weighted Ensemble Models Are Strong Continual Learners ECCV2024
In this work, we study the problem of continual learning (CL) where the goal is to learn a model on a sequence of tasks, such that the data from the previous tasks becomes unavailable while learning on the current task data. CL is essentially a balancing act between being able to learn on the new task (i.e., plasticity) and maintaining the performance on the previously learned concepts (i.e., stability). Intending to address the stability-plasticity trade-off, we propose to perform weight-ensembling of the model parameters of the previous and current tasks. This weighted-ensembled model, which we call Continual Model Averaging (or CoMA), attains high accuracy on the current task by leveraging plasticity, while not deviating too far from the previous weight configuration, ensuring stability. We also propose an improved variant of CoMA, named Continual Fisher-weighted Model Averaging (or CoFiMA), that selectively weighs each parameter in the weights ensemble by leveraging the Fisher information of the weights of the model. Both variants are conceptually simple, easy to implement, and effective in attaining state-of-the-art performance on several standard CL benchmarks. Code is available at: https://github.com/IemProg/CoFiMA.
comment: Accepted for ECCV2024, Code: https://github.com/IemProg/CoFiMA
♻ ☆ Extreme AutoML: Analysis of Classification, Regression, and NLP Performance
Utilizing machine learning techniques has always required choosing hyperparameters. This is true whether one uses a classical technique such as a KNN or very modern neural networks such as Deep Learning. Though in many applications, hyperparameters are chosen by hand, automated methods have become increasingly more common. These automated methods have become collectively known as automated machine learning, or AutoML. Several automated selection algorithms have shown similar or improved performance over state-of-the-art methods. This breakthrough has led to the development of cloud-based services like Google AutoML, which is based on Deep Learning and is widely considered to be the industry leader in AutoML services. Extreme Learning Machines (ELMs) use a fundamentally different type of neural architecture, producing better results at a significantly discounted computational cost. We benchmark the Extreme AutoML technology against Google's AutoML using several popular classification data sets from the University of California at Irvine's (UCI) repository, and several other data sets, observing significant advantages for Extreme AutoML in accuracy, Jaccard Indices, the variance of Jaccard Indices across classes (i.e. class variance) and training times.
comment: 8 pages, 7 figures
♻ ☆ Rethinking the Starting Point: Collaborative Pre-Training for Federated Downstream Tasks AAAI 2025
A few recent studies have demonstrated that leveraging centrally pre-trained models can offer advantageous initializations for federated learning (FL). However, existing pre-training methods do not generalize well when faced with an arbitrary set of downstream FL tasks. Specifically, they often (i) achieve limited average accuracy, particularly when there are unseen downstream labels, and (ii) result in significant accuracy variance, failing to provide a balanced performance across clients. To address these challenges, we propose CoPreFL, a collaborative/distributed pre-training approach which provides a robust initialization for downstream FL tasks. The key idea of CoPreFL is a model-agnostic meta-learning (MAML) procedure that tailors the global model to closely mimic heterogeneous and unseen FL scenarios, resulting in a pre-trained model that is rapidly adaptable to arbitrary FL tasks. Our MAML procedure incorporates performance variance into the meta-objective function, balancing performance across clients rather than solely optimizing for accuracy. Through extensive experiments, we demonstrate that CoPreFL obtains significant improvements in both average accuracy and variance across arbitrary downstream FL tasks with unseen/seen labels, compared with various pre-training baselines. We also show how CoPreFL is compatible with different well-known FL algorithms applied by the downstream tasks, enhancing performance in each case.
comment: AAAI 2025
♻ ☆ Sequence-Augmented SE(3)-Flow Matching For Conditional Protein Backbone Generation NeurIPS 2024
Proteins are essential for almost all biological processes and derive their diverse functions from complex 3D structures, which are in turn determined by their amino acid sequences. In this paper, we exploit the rich biological inductive bias of amino acid sequences and introduce FoldFlow-2, a novel sequence-conditioned SE(3)-equivariant flow matching model for protein structure generation. FoldFlow-2 presents substantial new architectural features over the previous FoldFlow family of models including a protein large language model to encode sequence, a new multi-modal fusion trunk that combines structure and sequence representations, and a geometric transformer based decoder. To increase diversity and novelty of generated samples -- crucial for de-novo drug design -- we train FoldFlow-2 at scale on a new dataset that is an order of magnitude larger than PDB datasets of prior works, containing both known proteins in PDB and high-quality synthetic structures achieved through filtering. We further demonstrate the ability to align FoldFlow-2 to arbitrary rewards, e.g. increasing secondary structures diversity, by introducing a Reinforced Finetuning (ReFT) objective. We empirically observe that FoldFlow-2 outperforms previous state-of-the-art protein structure-based generative models, improving over RFDiffusion in terms of unconditional generation across all metrics including designability, diversity, and novelty across all protein lengths, as well as exhibiting generalization on the task of equilibrium conformation sampling. Finally, we demonstrate that a fine-tuned FoldFlow-2 makes progress on challenging conditional design tasks such as designing scaffolds for the VHH nanobody.
comment: Presented at NeurIPS 2024
♻ ☆ Topic Classification of Case Law Using a Large Language Model and a New Taxonomy for UK Law: AI Insights into Summary Judgment
This paper addresses a critical gap in legal analytics by developing and applying a novel taxonomy for topic classification of summary judgment cases in the United Kingdom. Using a curated dataset of summary judgment cases, we use the Large Language Model Claude 3 Opus to explore functional topics and trends. We find that Claude 3 Opus correctly classified the topic with an accuracy of 87.13% and an F1 score of 0.87. The analysis reveals distinct patterns in the application of summary judgments across various legal domains. As case law in the United Kingdom is not originally labelled with keywords or a topic filtering option, the findings not only refine our understanding of the thematic underpinnings of summary judgments but also illustrate the potential of combining traditional and AI-driven approaches in legal classification. Therefore, this paper provides a new and general taxonomy for UK law. The implications of this work serve as a foundation for further research and policy discussions in the field of judicial administration and computational legal research methodologies.
♻ ☆ Inverting Gradient Attacks Makes Powerful Data Poisoning
Gradient attacks and data poisoning tamper with the training of machine learning algorithms to maliciously alter them and have been proven to be equivalent in convex settings. The extent of harm these attacks can produce in non-convex settings is still to be determined. Gradient attacks can affect far less systems than data poisoning but have been argued to be more harmful since they can be arbitrary, whereas data poisoning reduces the attacker's power to only being able to inject data points to training sets, via e.g. legitimate participation in a collaborative dataset. This raises the question of whether the harm made by gradient attacks can be matched by data poisoning in non-convex settings. In this work, we provide a positive answer in a worst-case scenario and show how data poisoning can mimic a gradient attack to perform an availability attack on (non-convex) neural networks. Through gradient inversion, commonly used to reconstruct data points from actual gradients, we show how reconstructing data points out of malicious gradients can be sufficient to perform a range of attacks. This allows us to show, for the first time, an availability attack on neural networks through data poisoning, that degrades the model's performances to random-level through a minority (as low as 1%) of poisoned points.
comment: 12 pages, 11 figures
♻ ☆ Dynamic Pricing and Advertising with Demand Learning
We consider a novel pricing and advertising framework, where a seller not only sets product price but also designs flexible 'advertising schemes' to influence customers' valuation of the product. We impose no structural restriction on the seller's feasible advertising strategies and allow her to advertise the product by disclosing or concealing any information. Following the literature in information design, this fully flexible advertising can be modeled as the seller being able to choose any information policy that signals the product quality/characteristic to the customers. Customers observe the advertising signal and infer a Bayesian belief over the products. We aim to investigate two questions in this work: (1) What is the value of advertising? To what extent can advertising enhance a seller's revenue? (2) Without any apriori knowledge of the customers' demand function, how can a seller adaptively learn and optimize both pricing and advertising strategies using past purchase responses? To study the first question, we introduce and study the value of advertising - a revenue gap between using advertising vs not advertising, and we provide a crisp tight characterization for this notion for a broad family of problems. For the second question, we study the seller's dynamic pricing and advertising problem with demand uncertainty. Our main result for this question is a computationally efficient online algorithm that achieves an optimal $O(T^{2/3}(m\log T)^{1/3})$ regret rate when the valuation function is linear in the product quality. Here $m$ is the cardinality of the discrete product quality domain and $T$ is the time horizon. This result requires some mild regularity assumptions on the valuation function, but no Lipschitz or smoothness assumption on the customers' demand function. We also obtain several improved results for the widely considered special case of additive valuations.
comment: Added new results, including a new section for detailed analysis of value of advertising, a section for numerical results. Also rewrite the introduction and setting section
♻ ☆ Shaping AI's Impact on Billions of Lives
Artificial Intelligence (AI), like any transformative technology, has the potential to be a double-edged sword, leading either toward significant advancements or detrimental outcomes for society as a whole. As is often the case when it comes to widely-used technologies in market economies (e.g., cars and semiconductor chips), commercial interest tends to be the predominant guiding factor. The AI community is at risk of becoming polarized to either take a laissez-faire attitude toward AI development, or to call for government overregulation. Between these two poles we argue for the community of AI practitioners to consciously and proactively work for the common good. This paper offers a blueprint for a new type of innovation infrastructure including 18 concrete milestones to guide AI research in that direction. Our view is that we are still in the early days of practical AI, and focused efforts by practitioners, policymakers, and other stakeholders can still maximize the upsides of AI and minimize its downsides. We talked to luminaries such as recent Nobelist John Jumper on science, President Barack Obama on governance, former UN Ambassador and former National Security Advisor Susan Rice on security, philanthropist Eric Schmidt on several topics, and science fiction novelist Neal Stephenson on entertainment. This ongoing dialogue and collaborative effort has produced a comprehensive, realistic view of what the actual impact of AI could be, from a diverse assembly of thinkers with deep understanding of this technology and these domains. From these exchanges, five recurring guidelines emerged, which form the cornerstone of a framework for beginning to harness AI in service of the public good. They not only guide our efforts in discovery but also shape our approach to deploying this transformative technology responsibly and ethically.
♻ ☆ Annotation-guided Protein Design with Multi-Level Domain Alignment KDD 2025
The core challenge of de novo protein design lies in creating proteins with specific functions or properties, guided by certain conditions. Current models explore to generate protein using structural and evolutionary guidance, which only provide indirect conditions concerning functions and properties. However, textual annotations of proteins, especially the annotations for protein domains, which directly describe the protein's high-level functionalities, properties, and their correlation with target amino acid sequences, remain unexplored in the context of protein design tasks. In this paper, we propose Protein-Annotation Alignment Generation, PAAG, a multi-modality protein design framework that integrates the textual annotations extracted from protein database for controllable generation in sequence space. Specifically, within a multi-level alignment module, PAAG can explicitly generate proteins containing specific domains conditioned on the corresponding domain annotations, and can even design novel proteins with flexible combinations of different kinds of annotations. Our experimental results underscore the superiority of the aligned protein representations from PAAG over 7 prediction tasks. Furthermore, PAAG demonstrates a significant increase in generation success rate (24.7% vs 4.7% in zinc finger, and 54.3% vs 22.0% in the immunoglobulin domain) in comparison to the existing model. We anticipate that PAAG will broaden the horizons of protein design by leveraging the knowledge from between textual annotation and proteins.
comment: Accepted by KDD 2025
♻ ☆ Robust Deep Reinforcement Learning Through Adversarial Attacks and Training : A Survey
Deep Reinforcement Learning (DRL) is a subfield of machine learning for training autonomous agents that take sequential actions across complex environments. Despite its significant performance in well-known environments, it remains susceptible to minor condition variations, raising concerns about its reliability in real-world applications. To improve usability, DRL must demonstrate trustworthiness and robustness. A way to improve the robustness of DRL to unknown changes in the environmental conditions and possible perturbations is through Adversarial Training, by training the agent against well-suited adversarial attacks on the observations and the dynamics of the environment. Addressing this critical issue, our work presents an in-depth analysis of contemporary adversarial attack and training methodologies, systematically categorizing them and comparing their objectives and operational mechanisms.
comment: 61 pages, 17 figues, 1 table
♻ ☆ Nonlinear optical encoding enabled by recurrent linear scattering
Optical information processing and computing can potentially offer enhanced performance, scalability and energy efficiency. However, achieving nonlinearity-a critical component of computation-remains challenging in the optical domain. Here we introduce a design that leverages a multiple-scattering cavity to passively induce optical nonlinear random mapping with a continuous-wave laser at a low power. Each scattering event effectively mixes information from different areas of a spatial light modulator, resulting in a highly nonlinear mapping between the input data and output pattern. We demonstrate that our design retains vital information even when the readout dimensionality is reduced, thereby enabling optical data compression. This capability allows our optical platforms to offer efficient optical information processing solutions across applications. We demonstrate our design's efficacy across tasks, including classification, image reconstruction, keypoint detection and object detection, all of which are achieved through optical data compression combined with a digital decoder. In particular, high performance at extreme compression ratios is observed in real-time pedestrian detection. Our findings open pathways for novel algorithms and unconventional architectural designs for optical computing.
comment: 14 pages, 4 figures
♻ ☆ Spatial-Temporal Federated Learning for Lifelong Person Re-identification on Distributed Edges
Data drift is a thorny challenge when deploying person re-identification (ReID) models into real-world devices, where the data distribution is significantly different from that of the training environment and keeps changing. To tackle this issue, we propose a federated spatial-temporal incremental learning approach, named FedSTIL, which leverages both lifelong learning and federated learning to continuously optimize models deployed on many distributed edge clients. Unlike previous efforts, FedSTIL aims to mine spatial-temporal correlations among the knowledge learnt from different edge clients. Specifically, the edge clients first periodically extract general representations of drifted data to optimize their local models. Then, the learnt knowledge from edge clients will be aggregated by centralized parameter server, where the knowledge will be selectively and attentively distilled from spatial- and temporal-dimension with carefully designed mechanisms. Finally, the distilled informative spatial-temporal knowledge will be sent back to correlated edge clients to further improve the recognition accuracy of each edge client with a lifelong learning method. Extensive experiments on a mixture of five real-world datasets demonstrate that our method outperforms others by nearly 4% in Rank-1 accuracy, while reducing communication cost by 62%. All implementation codes are publicly available on https://github.com/MSNLAB/Federated-Lifelong-Person-ReID
♻ ☆ LinVT: Empower Your Image-level Large Language Model to Understand Videos
Large Language Models (LLMs) have been widely used in various tasks, motivating us to develop an LLM-based assistant for videos. Instead of training from scratch, we propose a module to transform arbitrary well-trained image-based LLMs into video-LLMs (after being trained on video data). To better adapt image-LLMs for processing videos, we introduce two design principles: linear transformation to preserve the original visual-language alignment and representative information condensation from redundant video content. Guided by these principles, we propose a plug-and-play Linear Video Tokenizer(LinVT), which enables existing image-LLMs to understand videos. We benchmark LinVT with six recent visual LLMs: Aquila, Blip-3, InternVL2, Mipha, Molmo and Qwen2-VL, showcasing the high compatibility of LinVT. LinVT-based LLMs achieve state-of-the-art performance across various video benchmarks, illustrating the effectiveness of LinVT in multi-modal video understanding.
♻ ☆ Benchmarking Autoregressive Conditional Diffusion Models for Turbulent Flow Simulation
Simulating turbulent flows is crucial for a wide range of applications, and machine learning-based solvers are gaining increasing relevance. However, achieving temporal stability when generalizing to longer rollout horizons remains a persistent challenge for learned PDE solvers. In this work, we analyze if fully data-driven fluid solvers that utilize an autoregressive rollout based on conditional diffusion models are a viable option to address this challenge. We investigate accuracy, posterior sampling, spectral behavior, and temporal stability, while requiring that methods generalize to flow parameters beyond the training regime. To quantitatively and qualitatively benchmark the performance of various flow prediction approaches, three challenging 2D scenarios including incompressible and transonic flows, as well as isotropic turbulence are employed. We find that even simple diffusion-based approaches can outperform multiple established flow prediction methods in terms of accuracy and temporal stability, while being on par with state-of-the-art stabilization techniques like unrolling at training time. Such traditional architectures are superior in terms of inference speed, however, the probabilistic nature of diffusion approaches allows for inferring multiple predictions that align with the statistics of the underlying physics. Overall, our benchmark contains three carefully chosen data sets that are suitable for probabilistic evaluation alongside various established flow prediction architectures.
comment: Source code available at https://github.com/tum-pbs/autoreg-pde-diffusion and further information and videos at https://ge.in.tum.de/publications/2023-acdm-kohl
♻ ☆ TRIGS: Trojan Identification from Gradient-based Signatures
Training machine learning models can be very expensive or even unaffordable. This may be, for example, due to data limitations, such as unavailability or being too large, or computational power limitations. Therefore, it is a common practice to rely on open-source pre-trained models whenever possible.However, this practice is alarming from a security perspective. Pre-trained models can be infected with Trojan attacks, in which the attacker embeds a trigger in the model such that the model's behavior can be controlled by the attacker when the trigger is present in the input. In this paper, we present a novel method for detecting Trojan models. Our method creates a signature for a model based on activation optimization. A classifier is then trained to detect a Trojan model given its signature. We call our method TRIGS for TRojan Identification from Gradient-based Signatures. TRIGS achieves state-of-the-art performance on two public datasets of convolutional models. Additionally, we introduce a new challenging dataset of ImageNet models based on the vision transformer architecture. TRIGS delivers the best performance on the new dataset, surpassing the baseline methods by a large margin. Our experiments also show that TRIGS requires only a small amount of clean samples to achieve good performance, and works reasonably well even if the defender does not have prior knowledge about the attacker's model architecture. Our code and data are publicly available.
♻ ☆ A Robust Clustering Framework Combining Minimum Description Length and Genetic Optimization
Clustering algorithms are fundamental in data analysis, enabling the organization of data into meaningful groups. However, individual clustering methods often face limitations and biases, making it challenging to develop a universal solution for diverse datasets. To address this, we propose a novel clustering framework that combines the Minimum Description Length (MDL) principle with a genetic optimization algorithm. This approach begins with an ensemble clustering solution as a baseline, which is refined using MDL-based evaluation functions and optimized with a genetic algorithm. By leveraging the MDL principle, the method adapts to the intrinsic properties of datasets, minimizing dependence on input clusters and ensuring a data-driven process. The proposed method was evaluated on thirteen benchmark datasets using four validation metrics: accuracy, normalized mutual information (NMI), Fisher score, and adjusted Rand index (ARI). Results show that the method consistently outperforms traditional clustering algorithms, achieving higher accuracy, greater stability, and reduced biases. Its adaptability makes it a reliable tool for clustering complex and varied datasets. This study demonstrates the potential of combining MDL and genetic optimization to create a robust and versatile clustering framework, advancing the field of data analysis and offering a scalable solution for diverse applications.
♻ ☆ Assessing the Impact of Image Dataset Features on Privacy-Preserving Machine Learning
Machine Learning (ML) is crucial in many sectors, including computer vision. However, ML models trained on sensitive data face security challenges, as they can be attacked and leak information. Privacy-Preserving Machine Learning (PPML) addresses this by using Differential Privacy (DP) to balance utility and privacy. This study identifies image dataset characteristics that affect the utility and vulnerability of private and non-private Convolutional Neural Network (CNN) models. Through analyzing multiple datasets and privacy budgets, we find that imbalanced datasets increase vulnerability in minority classes, but DP mitigates this issue. Datasets with fewer classes improve both model utility and privacy, while high entropy or low Fisher Discriminant Ratio (FDR) datasets deteriorate the utility-privacy trade-off. These insights offer valuable guidance for practitioners and researchers in estimating and optimizing the utility-privacy trade-off in image datasets, helping to inform data and privacy modifications for better outcomes based on dataset characteristics.
comment: Accepted at 21st Conference on Database Systems for Business, Technology and Web (BTW 2025)
♻ ☆ Dual-Directed Algorithm Design for Efficient Pure Exploration COLT'23
We consider pure-exploration problems in the context of stochastic sequential adaptive experiments with a finite set of alternatives. The central objective is to answer a query regarding the alternatives with high confidence while minimizing measurement efforts. One canonical example is identifying the best-performing alternative, a problem known as ranking and selection in simulation or best-arm identification in machine learning. We formulate the problem complexity measure as a maximin optimization problem for the static fixed-budget, fixed-confidence, and posterior convergence rate settings. By incorporating dual variables directly into the analysis, we derive necessary and sufficient conditions for an allocation's optimality. The introduction of dual variables allows us to sidestep the combinatorial complexity that arises when considering only primal variables. These optimality conditions enable the extension of the top-two algorithm design principle to more general pure-exploration problems. Moreover, our analysis yields a straightforward and effective information-directed selection rule that adaptively chooses from a candidate set based on the informational value of the candidates. We demonstrate the broad range of contexts in which our design principle can be implemented. In particular, when combined with information-directed selection, top-two Thompson sampling achieves asymptotic optimality in Gaussian best-arm identification, resolving a notable open question in the pure-exploration literature. Our algorithm attains optimality in $\varepsilon$-best-arm identification (or ranking and selection with a probability of good selection guarantee) and thresholding bandits. Our results provide a general principle for adapting Thompson sampling to general pure-exploration problems. Numerical experiments highlight the efficiency of our proposed algorithms compared to existing methods.
comment: An earlier version of this paper appeared as an extended abstract in the Proceedings of the 36th Annual Conference on Learning Theory, COLT'23, with the title "Information-Directed Selection for Top-Two Algorithms.''
♻ ☆ Graph Agent Network: Empowering Nodes with Inference Capabilities for Adversarial Resilience
End-to-end training with global optimization have popularized graph neural networks (GNNs) for node classification, yet inadvertently introduced vulnerabilities to adversarial edge-perturbing attacks. Adversaries can exploit the inherent opened interfaces of GNNs' input and output, perturbing critical edges and thus manipulating the classification results. Current defenses, due to their persistent utilization of global-optimization-based end-to-end training schemes, inherently encapsulate the vulnerabilities of GNNs. This is specifically evidenced in their inability to defend against targeted secondary attacks. In this paper, we propose the Graph Agent Network (GAgN) to address the aforementioned vulnerabilities of GNNs. GAgN is a graph-structured agent network in which each node is designed as an 1-hop-view agent. Through the decentralized interactions between agents, they can learn to infer global perceptions to perform tasks including inferring embeddings, degrees and neighbor relationships for given nodes. This empowers nodes to filtering adversarial edges while carrying out classification tasks. Furthermore, agents' limited view prevents malicious messages from propagating globally in GAgN, thereby resisting global-optimization-based secondary attacks. We prove that single-hidden-layer multilayer perceptrons (MLPs) are theoretically sufficient to achieve these functionalities. Experimental results show that GAgN effectively implements all its intended capabilities and, compared to state-of-the-art defenses, achieves optimal classification accuracy on the perturbed datasets.
♻ ☆ MT2ST: Adaptive Multi-Task to Single-Task Learning
Efficient machine learning (ML) has become increasingly important as models grow larger and data volumes expand. In this work, we address the trade-off between generalization in multi-task learning (MTL) and precision in single-task learning (STL) by introducing the Multi-Task to Single-Task (MT2ST) framework. MT2ST is designed to enhance training efficiency and accuracy in word embedding tasks, showcasing its value as a practical application of efficient ML. Our framework employs two strategies: *Diminish*, which gradually reduces the influence of auxiliary tasks, and *Switch*, which transitions training from MTL to STL at a specific point. Empirical results show that MT2ST reduces training time by 67\% compared to STL and by 13\% compared to traditional MTL, while maintaining high accuracy. These findings highlight MT2ST as an efficient ML solution tailored for optimizing word embedding training. Code is available at https://github.com/NoakLiu/MT2ST.
♻ ☆ Understanding Token Probability Encoding in Output Embeddings COLING 2025
In this paper, we investigate the output token probability information in the output embedding of language models. We find an approximate common log-linear encoding of output token probabilities within the output embedding vectors and empirically demonstrate that it is accurate and sparse. As a causality examination, we steer the encoding in output embedding to modify the output probability distribution accurately. Moreover, the sparsity we find in output probability encoding suggests that a large number of dimensions in the output embedding do not contribute to causal language modeling. Therefore, we attempt to delete the output-unrelated dimensions and find more than 30% of the dimensions can be deleted without significant movement in output distribution and sequence generation. Additionally, in the pre-training dynamics of language models, we find that the output embeddings capture the corpus token frequency information in early steps, even before an obvious convergence of parameters starts.
comment: 15 pages, 17 figures, 3 tables. COLING 2025 Accepted
Rate-In: Information-Driven Adaptive Dropout Rates for Improved Inference-Time Uncertainty Estimation
Accurate uncertainty estimation is crucial for deploying neural networks in risk-sensitive applications such as medical diagnosis. Monte Carlo Dropout is a widely used technique for approximating predictive uncertainty by performing stochastic forward passes with dropout during inference. However, using static dropout rates across all layers and inputs can lead to suboptimal uncertainty estimates, as it fails to adapt to the varying characteristics of individual inputs and network layers. Existing approaches optimize dropout rates during training using labeled data, resulting in fixed inference-time parameters that cannot adjust to new data distributions, compromising uncertainty estimates in Monte Carlo simulations. In this paper, we propose Rate-In, an algorithm that dynamically adjusts dropout rates during inference by quantifying the information loss induced by dropout in each layer's feature maps. By treating dropout as controlled noise injection and leveraging information-theoretic principles, Rate-In adapts dropout rates per layer and per input instance without requiring ground truth labels. By quantifying the functional information loss in feature maps, we adaptively tune dropout rates to maintain perceptual quality across diverse medical imaging tasks and architectural configurations. Our extensive empirical study on synthetic data and real-world medical imaging tasks demonstrates that Rate-In improves calibration and sharpens uncertainty estimates compared to fixed or heuristic dropout rates without compromising predictive performance. Rate-In offers a practical, unsupervised, inference-time approach to optimizing dropout for more reliable predictive uncertainty estimation in critical applications.
comment: Updated author affiliation
♻ ☆ Learning incomplete factorization preconditioners for GMRES
Incomplete LU factorizations of sparse matrices are widely used as preconditioners in Krylov subspace methods to speed up solving linear systems. Unfortunately, computing the preconditioner itself can be time-consuming and sensitive to hyper-parameters. Instead, we replace the hand-engineered algorithm with a graph neural network that is trained to approximate the matrix factorization directly. To apply the output of the neural network as a preconditioner, we propose an output activation function that guarantees that the predicted factorization is invertible. Further, applying a graph neural network architecture allows us to ensure that the output itself is sparse which is desirable from a computational standpoint. We theoretically analyze and empirically evaluate different loss functions to train the learned preconditioners and show their effectiveness in decreasing the number of GMRES iterations and improving the spectral properties on synthetic data. The code is available at https://github.com/paulhausner/neural-incomplete-factorization.
comment: The first two authors contributed equally, Northern Lights Deep Learning Conference, 15 pages
♻ ☆ Global and Preference-based Optimization with Mixed Variables using Piecewise Affine Surrogates
Optimization problems involving mixed variables (i.e., variables of numerical and categorical nature) can be challenging to solve, especially in the presence of mixed-variable constraints. Moreover, when the objective function is the result of a complicated simulation or experiment, it may be expensive-to-evaluate. This paper proposes a novel surrogate-based global optimization algorithm to solve linearly constrained mixed-variable problems up to medium size (around 100 variables after encoding). The proposed approach is based on constructing a piecewise affine surrogate of the objective function over feasible samples. We assume the objective function is black-box and expensive-to-evaluate, while the linear constraints are quantifiable, unrelaxable, a priori known, and are cheap to evaluate. We introduce two types of exploration functions to efficiently search the feasible domain via mixed-integer linear programming solvers. We also provide a preference-based version of the algorithm designed for situations where only pairwise comparisons between samples can be acquired, while the underlying objective function to minimize remains unquantified. The two algorithms are evaluated on several unconstrained and constrained mixed-variable benchmark problems. The results show that, within a small number of required experiments/simulations, the proposed algorithms can often achieve better or comparable results than other existing methods.
comment: code available at https://github.com/mjzhu-p/PWAS
♻ ☆ Uncovering Capabilities of Model Pruning in Graph Contrastive Learning
Graph contrastive learning has achieved great success in pre-training graph neural networks without ground-truth labels. Leading graph contrastive learning follows the classical scheme of contrastive learning, forcing model to identify the essential information from augmented views. However, general augmented views are produced via random corruption or learning, which inevitably leads to semantics alteration. Although domain knowledge guided augmentations alleviate this issue, the generated views are domain specific and undermine the generalization. In this work, motivated by the firm representation ability of sparse model from pruning, we reformulate the problem of graph contrastive learning via contrasting different model versions rather than augmented views. We first theoretically reveal the superiority of model pruning in contrast to data augmentations. In practice, we take original graph as input and dynamically generate a perturbed graph encoder to contrast with the original encoder by pruning its transformation weights. Furthermore, considering the integrity of node embedding in our method, we are capable of developing a local contrastive loss to tackle the hard negative samples that disturb the model training. We extensively validate our method on various benchmarks regarding graph classification via unsupervised and transfer learning. Compared to the state-of-the-art (SOTA) works, better performance can always be obtained by the proposed method.
comment: MM' 24
♻ ☆ Active learning for efficient discovery of optimal gene combinations in the combinatorial perturbation space
The advancement of novel combinatorial CRISPR screening technologies enables the identification of synergistic gene combinations on a large scale. This is crucial for developing novel and effective combination therapies, but the combinatorial space makes exhaustive experimentation infeasible. We introduce NAIAD, an active learning framework that efficiently discovers optimal gene pairs capable of driving cells toward desired cellular phenotypes. NAIAD leverages single-gene perturbation effects and adaptive gene embeddings that scale with the training data size, mitigating overfitting in small-sample learning while capturing complex gene interactions as more data is collected. Evaluated on four CRISPR combinatorial perturbation datasets totaling over 350,000 genetic interactions, NAIAD, trained on small datasets, outperforms existing models by up to 40\% relative to the second-best. NAIAD's recommendation system prioritizes gene pairs with the maximum predicted effects, resulting in the highest marginal gain in each AI-experiment round and accelerating discovery with fewer CRISPR experimental iterations. Our NAIAD framework (https://github.com/NeptuneBio/NAIAD) improves the identification of novel, effective gene combinations, enabling more efficient CRISPR library design and offering promising applications in genomics research and therapeutic development.
♻ ☆ DexHandDiff: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simpler manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexHandDiff, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexHandDiff models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, and hammer striking demonstrate DexHandDiff's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves 70.0% success on 30-degree door opening, 40.0% and 36.7% on pen and block half-side re-orientation respectively, and 46.7% on hammer nail half drive, highlighting its robustness and flexibility in contact-rich manipulation.
comment: 27 pages (new name). Project page: https://dexdiffuser.github.io/
♻ ☆ Efficient Large Foundation Model Inference: A Perspective From Model and System Co-Design
As Large Language Models (LLMs) become popular, the need for efficient design for ML models on LLMs grows. We are amazed by the excellent output by the LLMs, yet we are still troubled with slow inference speed and large memory consumption of contemporary LLMs. This paper focuses on modern efficient inference technologies on LLMs and illustrates them from two perspectives: model and system design. These methodologies optimize LLM inference from different aspects to save computational resources, making LLMs more efficient, affordable, and more accessible.
♻ ☆ FTFT: Efficient and Robust Fine-Tuning by Transferring Training Dynamics COLING 2025
Despite the massive success of fine-tuning Pre-trained Language Models (PLMs), they remain susceptible to out-of-distribution input. Dataset cartography is a simple yet effective dual-model approach that improves the robustness of fine-tuned PLMs. It involves fine-tuning a model on the original training set (i.e. reference model), selecting a subset of important training instances based on the training dynamics, and fine-tuning again only on these selected examples (i.e. main model). However, this approach requires fine-tuning the same model twice, which is computationally expensive for large PLMs. In this paper, we show that (1) training dynamics are highly transferable across model sizes and pre-training methods, and that (2) fine-tuning main models using these selected training instances achieves higher training efficiency than empirical risk minimization (ERM). Building on these observations, we propose a novel fine-tuning approach: Fine-Tuning by transFerring Training dynamics (FTFT). Compared with dataset cartography, FTFT uses more efficient reference models and aggressive early stopping. FTFT achieves robustness improvements over ERM while lowering the training cost by up to $\sim 50\%$.
comment: COLING 2025 Camera-Ready
♻ ☆ Domain-Adaptive Pre-training of Self-Supervised Foundation Models for Medical Image Classification in Gastrointestinal Endoscopy
Video capsule endoscopy has transformed gastrointestinal endoscopy (GIE) diagnostics by offering a non-invasive method for capturing detailed images of the gastrointestinal tract, enabling early disease detection. However, its potential is limited by the sheer volume of images generated during the imaging procedure, which can take anywhere from 6-8 hours and often produce up to 1 million images, necessitating automated analysis. Additionally, the variability of these images, combined with the need for expert annotations and the scarcity of large, high-quality labeled datasets, constrains the effectiveness of current medical image analysis models. To address this, we introduce a novel large GIE dataset, called EndoExtend24, created by merging ten existing public and private datasets, ensuring patient integrity across splits. EndoExtend24 includes over 226,000 labeled images, as well as dynamic class mappings, which allow unified training across datasets with differing labeling granularity, supporting up to 123 distinct pathological findings. Further, we propose to leverage domain adaptive pre-training of foundation models trained with self-supervision on generic image data, to adapt them to the task of GIE medical image diagnosis. Specifically, the EVA-02 model, which is based on the ViT architecture and trained on ImageNet-22k with masked image modeling (using EVA-CLIP as a MIM teacher), is pre-trained on the EndoExtend24 dataset to achieve domain adaptation, and finally trained on the Capsule Endoscopy 2024 Challenge dataset. Our model demonstrates robust performance, securing third place in the Capsule Endoscopy 2024 Challenge. We achieved a macro AUC of 0.762 and a balanced accuracy of 37.1% on the test set. These results emphasize the effectiveness of our domain-adaptive pre-training approach and the enriched EndoExtend24 dataset in advancing gastrointestinal endoscopy diagnostics.
♻ ☆ Which Augmentation Should I Use? An Empirical Investigation of Augmentations for Self-Supervised Phonocardiogram Representation Learning
Despite recent advancements in deep learning, its application in real-world medical settings, such as phonocardiogram (PCG) classification, remains limited. A significant barrier is the lack of high-quality annotated datasets, which hampers the development of robust, generalizable models that can perform well on newly collected, out-of-distribution (OOD) data. Self-Supervised Learning (SSL) contrastive learning, has shown promise in mitigating the issue of data scarcity by using unlabeled data to enhance model robustness. Even though SSL methods have been proposed and researched in other domains, works focusing on the impact of data augmentations on model robustness for PCG classification are limited. In particular, while augmentations are a key component in SSL, selecting the most suitable policy during training is highly challenging. Improper augmentations can lead to substantial performance degradation and even hinder a network's ability to learn meaningful representations. Addressing this gap, our research aims to explore and evaluate a wide range of audio-based augmentations and uncover combinations that enhance SSL model performance in PCG classification. We conduct a comprehensive comparative analysis across multiple datasets, assessing the impact of various augmentations on model performance. Our findings reveal that depending on the training distribution, augmentation choice significantly influences model robustness, with fully-supervised models experiencing up to a 32\% drop in effectiveness when evaluated on unseen data, while SSL models demonstrate greater resilience, losing only 10\% or even improving in some cases. This study also highlights the most promising and appropriate augmentations for PCG signal processing, by calculating their effect size on training. These insights equip researchers with valuable guidelines for developing reliable models in PCG signal processing.
comment: Accepted in IEEE ACCESS
♻ ☆ Improving Deep Reinforcement Learning by Reducing the Chain Effect of Value and Policy Churn NeurIPS 2024
Deep neural networks provide Reinforcement Learning (RL) powerful function approximators to address large-scale decision-making problems. However, these approximators introduce challenges due to the non-stationary nature of RL training. One source of the challenges in RL is that output predictions can churn, leading to uncontrolled changes after each batch update for states not included in the batch. Although such a churn phenomenon exists in each step of network training, how churn occurs and impacts RL remains under-explored. In this work, we start by characterizing churn in a view of Generalized Policy Iteration with function approximation, and we discover a chain effect of churn that leads to a cycle where the churns in value estimation and policy improvement compound and bias the learning dynamics throughout the iteration. Further, we concretize the study and focus on the learning issues caused by the chain effect in different settings, including greedy action deviation in value-based methods, trust region violation in proximal policy optimization, and dual bias of policy value in actor-critic methods. We then propose a method to reduce the chain effect across different settings, called Churn Approximated ReductIoN (CHAIN), which can be easily plugged into most existing DRL algorithms. Our experiments demonstrate the effectiveness of our method in both reducing churn and improving learning performance across online and offline, value-based and policy-based RL settings, as well as a scaling setting.
comment: Accepted to NeurIPS 2024. Project page: https://bluecontra.github.io/CHAIN
♻ ☆ Understanding Gradient Descent through the Training Jacobian
We examine the geometry of neural network training using the Jacobian of trained network parameters with respect to their initial values. Our analysis reveals low-dimensional structure in the training process which is dependent on the input data but largely independent of the labels. We find that the singular value spectrum of the Jacobian matrix consists of three distinctive regions: a "chaotic" region of values orders of magnitude greater than one, a large "bulk" region of values extremely close to one, and a "stable" region of values less than one. Along each bulk direction, the left and right singular vectors are nearly identical, indicating that perturbations to the initialization are carried through training almost unchanged. These perturbations have virtually no effect on the network's output in-distribution, yet do have an effect far out-of-distribution. While the Jacobian applies only locally around a single initialization, we find substantial overlap in bulk subspaces for different random seeds. Our code is available at https://github.com/EleutherAI/training-jacobian
♻ ☆ Mjolnir: Breaking the Shield of Perturbation-Protected Gradients via Adaptive Diffusion AAAI 2025
Perturbation-based mechanisms, such as differential privacy, mitigate gradient leakage attacks by introducing noise into the gradients, thereby preventing attackers from reconstructing clients' private data from the leaked gradients. However, can gradient perturbation protection mechanisms truly defend against all gradient leakage attacks? In this paper, we present the first attempt to break the shield of gradient perturbation protection in Federated Learning for the extraction of private information. We focus on common noise distributions, specifically Gaussian and Laplace, and apply our approach to DNN and CNN models. We introduce Mjolnir, a perturbation-resilient gradient leakage attack that is capable of removing perturbations from gradients without requiring additional access to the original model structure or external data. Specifically, we leverage the inherent diffusion properties of gradient perturbation protection to develop a novel diffusion-based gradient denoising model for Mjolnir. By constructing a surrogate client model that captures the structure of perturbed gradients, we obtain crucial gradient data for training the diffusion model. We further utilize the insight that monitoring disturbance levels during the reverse diffusion process can enhance gradient denoising capabilities, allowing Mjolnir to generate gradients that closely approximate the original, unperturbed versions through adaptive sampling steps. Extensive experiments demonstrate that Mjolnir effectively recovers the protected gradients and exposes the Federated Learning process to the threat of gradient leakage, achieving superior performance in gradient denoising and private data recovery.
comment: Accepted by AAAI 2025
♻ ☆ DGNN-YOLO: Interpretable Dynamic Graph Neural Networks with YOLO11 for Small Object Detection and Tracking in Traffic Surveillance
Accurate detection and tracking of small objects, such as pedestrians, cyclists, and motorbikes, is critical for traffic surveillance systems, which are crucial for improving road safety and decision-making in intelligent transportation systems. However, traditional methods face challenges such as occlusion, low resolution, and dynamic traffic conditions, necessitating innovative approaches to address these limitations. This paper introduces DGNN-YOLO, a novel framework integrating dynamic graph neural networks (DGNN) with YOLO11 to enhance small-object detection and tracking in traffic surveillance systems. The framework leverages YOLO11's advanced spatial feature extraction capabilities for precise object detection and incorporates a DGNN to model spatial-temporal relationships for robust real-time tracking dynamically. By constructing and updating graph structures, DGNN-YOLO effectively represents objects as nodes and their interactions as edges, thereby ensuring adaptive and accurate tracking in complex and dynamic environments. Additionally, Grad-CAM, Grad-CAM++, and Eigen-CAM visualization techniques were applied to DGNN-YOLO to provide model-agnostic interpretability and deeper insights into the model's decision-making process, enhancing its transparency and trustworthiness. Extensive experiments demonstrated that DGNN-YOLO consistently outperformed state-of-the-art methods in detecting and tracking small objects under diverse traffic conditions, achieving the highest precision (0.8382), recall (0.6875), and mAP@0.5:0.95 (0.6476), showing its robustness and scalability, particularly in challenging scenarios involving small and occluded objects. This study provides a scalable, real-time traffic surveillance and analysis solution, significantly contributing to intelligent transportation systems.
♻ ☆ Language Ranker: A Metric for Quantifying LLM Performance Across High and Low-Resource Languages AAAI 2025
The development of Large Language Models (LLMs) relies on extensive text corpora, which are often unevenly distributed across languages. This imbalance results in LLMs performing significantly better on high-resource languages like English, German, and French, while their capabilities in low-resource languages remain inadequate. Currently, there is a lack of quantitative methods to evaluate the performance of LLMs in these low-resource languages. To address this gap, we propose the Language Ranker, an intrinsic metric designed to benchmark and rank languages based on LLM performance using internal representations. By comparing the LLM's internal representation of various languages against a baseline derived from English, we can assess the model's multilingual capabilities in a robust and language-agnostic manner. Our analysis reveals that high-resource languages exhibit higher similarity scores with English, demonstrating superior performance, while low-resource languages show lower similarity scores, underscoring the effectiveness of our metric in assessing language-specific capabilities. Besides, the experiments show that there is a strong correlation between the LLM's performance in different languages and the proportion of those languages in its pre-training corpus. These insights underscore the efficacy of the Language Ranker as a tool for evaluating LLM performance across different languages, particularly those with limited resources.
comment: Accepted by AAAI 2025 (Social Impact Track)
♻ ☆ Hidden in the Noise: Two-Stage Robust Watermarking for Images
As the quality of image generators continues to improve, deepfakes become a topic of considerable societal debate. Image watermarking allows responsible model owners to detect and label their AI-generated content, which can mitigate the harm. Yet, current state-of-the-art methods in image watermarking remain vulnerable to forgery and removal attacks. This vulnerability occurs in part because watermarks distort the distribution of generated images, unintentionally revealing information about the watermarking techniques. In this work, we first demonstrate a distortion-free watermarking method for images, based on a diffusion model's initial noise. However, detecting the watermark requires comparing the initial noise reconstructed for an image to all previously used initial noises. To mitigate these issues, we propose a two-stage watermarking framework for efficient detection. During generation, we augment the initial noise with generated Fourier patterns to embed information about the group of initial noises we used. For detection, we (i) retrieve the relevant group of noises, and (ii) search within the given group for an initial noise that might match our image. This watermarking approach achieves state-of-the-art robustness to forgery and removal against a large battery of attacks.
♻ ☆ Towards Data-centric Machine Learning on Directed Graphs: a Survey
In recent years, Graph Neural Networks (GNNs) have made significant advances in processing structured data. However, most of them primarily adopted a model-centric approach, which simplifies graphs by converting them into undirected formats and emphasizes model designs. This approach is inherently limited in real-world applications due to the unavoidable information loss in simple undirected graphs and the model optimization challenges that arise when exceeding the upper bounds of this sub-optimal data representational capacity. As a result, there has been a shift toward data-centric methods that prioritize improving graph quality and representation. Specifically, various types of graphs can be derived from naturally structured data, including heterogeneous graphs, hypergraphs, and directed graphs. Among these, directed graphs offer distinct advantages in topological systems by modeling causal relationships, and directed GNNs have been extensively studied in recent years. However, a comprehensive survey of this emerging topic is still lacking. Therefore, we aim to provide a comprehensive review of directed graph learning, with a particular focus on a data-centric perspective. Specifically, we first introduce a novel taxonomy for existing studies. Subsequently, we re-examine these methods from the data-centric perspective, with an emphasis on understanding and improving data representation. It demonstrates that a deep understanding of directed graphs and their quality plays a crucial role in model performance. Additionally, we explore the diverse applications of directed GNNs across 10+ domains, highlighting their broad applicability. Finally, we identify key opportunities and challenges within the field, offering insights that can guide future research and development in directed graph learning.
comment: In Progress
♻ ☆ Efficient Prompt Tuning by Multi-Space Projection and Prompt Fusion
Prompt tuning is a promising method to fine-tune a pre-trained language model without retraining its large-scale parameters. Instead, it attaches a soft prompt to the input text, whereby downstream tasks can be well adapted by merely learning the embeddings of prompt tokens. Nevertheless, existing methods still suffer from two challenges: (i) they are hard to balance accuracy and efficiency. A longer (shorter) soft prompt generally leads to a better(worse) accuracy but at the cost of more (less) training time. (ii)The performance may not be consistent when adapting to different downstream tasks. We attribute it to the same embedding space but responsible for different requirements of downstream tasks. To address these issues, we propose an Efficient Prompt Tuning method (EPT) by multi-space projection and prompt fusion. Specifically, it decomposes a given soft prompt into a shorter prompt and two low-rank matrices, significantly reducing the training time. Accuracy is also enhanced by leveraging low-rank matrices and the short prompt as additional knowledge sources to enrich the semantics of the original short prompt. In addition, we project the soft prompt into multiple subspaces to improve the performance consistency, and then adaptively learn the combination weights of different spaces through a gating network. Experiments on 13 natural language processing downstream tasks show that our method significantly and consistently outperforms 11 comparison methods with the relative percentage of improvements up to 12.9%, and training time decreased by 14%.
♻ ☆ Scaling Up Probabilistic Circuits by Latent Variable Distillation
Probabilistic Circuits (PCs) are a unified framework for tractable probabilistic models that support efficient computation of various probabilistic queries (e.g., marginal probabilities). One key challenge is to scale PCs to model large and high-dimensional real-world datasets: we observe that as the number of parameters in PCs increases, their performance immediately plateaus. This phenomenon suggests that the existing optimizers fail to exploit the full expressive power of large PCs. We propose to overcome such bottleneck by latent variable distillation: we leverage the less tractable but more expressive deep generative models to provide extra supervision over the latent variables of PCs. Specifically, we extract information from Transformer-based generative models to assign values to latent variables of PCs, providing guidance to PC optimizers. Experiments on both image and language modeling benchmarks (e.g., ImageNet and WikiText-2) show that latent variable distillation substantially boosts the performance of large PCs compared to their counterparts without latent variable distillation. In particular, on the image modeling benchmarks, PCs achieve competitive performance against some of the widely-used deep generative models, including variational autoencoders and flow-based models, opening up new avenues for tractable generative modeling. Our code can be found at https://github.com/UCLA-StarAI/LVD.
♻ ☆ Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space NeurIPS 2024
Modern generative models demonstrate impressive capabilities, likely stemming from an ability to identify and manipulate abstract concepts underlying their training data. However, fundamental questions remain: what determines the concepts a model learns, the order in which it learns them, and its ability to manipulate those concepts? To address these questions, we propose analyzing a model's learning dynamics via a framework we call the concept space, where each axis represents an independent concept underlying the data generating process. By characterizing learning dynamics in this space, we identify how the speed at which a concept is learned, and hence the order of concept learning, is controlled by properties of the data we term concept signal. Further, we observe moments of sudden turns in the direction of a model's learning dynamics in concept space. Surprisingly, these points precisely correspond to the emergence of hidden capabilities, i.e., where latent interventions show the model possesses the capability to manipulate a concept, but these capabilities cannot yet be elicited via naive input prompting. While our results focus on synthetically defined toy datasets, we hypothesize a general claim on emergence of hidden capabilities may hold: generative models possess latent capabilities that emerge suddenly and consistently during training, though a model might not exhibit these capabilities under naive input prompting.
comment: NeurIPS 2024 (Spotlight)
♻ ☆ Image Inpainting via Tractable Steering of Diffusion Models
Diffusion models are the current state of the art for generating photorealistic images. Controlling the sampling process for constrained image generation tasks such as inpainting, however, remains challenging since exact conditioning on such constraints is intractable. While existing methods use various techniques to approximate the constrained posterior, this paper proposes to exploit the ability of Tractable Probabilistic Models (TPMs) to exactly and efficiently compute the constrained posterior, and to leverage this signal to steer the denoising process of diffusion models. Specifically, this paper adopts a class of expressive TPMs termed Probabilistic Circuits (PCs). Building upon prior advances, we further scale up PCs and make them capable of guiding the image generation process of diffusion models. Empirical results suggest that our approach can consistently improve the overall quality and semantic coherence of inpainted images across three natural image datasets (i.e., CelebA-HQ, ImageNet, and LSUN) with only $\sim\! 10 \%$ additional computational overhead brought by the TPM. Further, with the help of an image encoder and decoder, our method can readily accept semantic constraints on specific regions of the image, which opens up the potential for more controlled image generation tasks. In addition to proposing a new framework for constrained image generation, this paper highlights the benefit of more tractable models and motivates the development of expressive TPMs.
♻ ☆ Mixture of Online and Offline Experts for Non-stationary Time Series
We consider a general and realistic scenario involving non-stationary time series, consisting of several offline intervals with different distributions within a fixed offline time horizon, and an online interval that continuously receives new samples. For non-stationary time series, the data distribution in the current online interval may have appeared in previous offline intervals. We theoretically explore the feasibility of applying knowledge from offline intervals to the current online interval. To this end, we propose the Mixture of Online and Offline Experts (MOOE). MOOE learns static offline experts from offline intervals and maintains a dynamic online expert for the current online interval. It then adaptively combines the offline and online experts using a meta expert to make predictions for the samples received in the online interval. Specifically, we focus on theoretical analysis, deriving parameter convergence, regret bounds, and generalization error bounds to prove the effectiveness of the algorithm.
♻ ☆ CohEx: A Generalized Framework for Cohort Explanation
eXplainable Artificial Intelligence (XAI) has garnered significant attention for enhancing transparency and trust in machine learning models. However, the scopes of most existing explanation techniques focus either on offering a holistic view of the explainee model (global explanation) or on individual instances (local explanation), while the middle ground, i.e., cohort-based explanation, is less explored. Cohort explanations offer insights into the explainee's behavior on a specific group or cohort of instances, enabling a deeper understanding of model decisions within a defined context. In this paper, we discuss the unique challenges and opportunities associated with measuring cohort explanations, define their desired properties, and create a generalized framework for generating cohort explanations based on supervised clustering.
♻ ☆ NushuRescue: Revitalization of the Endangered Nushu Language with AI COLING 2025
The preservation and revitalization of endangered and extinct languages is a meaningful endeavor, conserving cultural heritage while enriching fields like linguistics and anthropology. However, these languages are typically low-resource, making their reconstruction labor-intensive and costly. This challenge is exemplified by Nushu, a rare script historically used by Yao women in China for self-expression within a patriarchal society. To address this challenge, we introduce NushuRescue, an AI-driven framework designed to train large language models (LLMs) on endangered languages with minimal data. NushuRescue automates evaluation and expands target corpora to accelerate linguistic revitalization. As a foundational component, we developed NCGold, a 500-sentence Nushu-Chinese parallel corpus, the first publicly available dataset of its kind. Leveraging GPT-4-Turbo, with no prior exposure to Nushu and only 35 short examples from NCGold, NushuRescue achieved 48.69% translation accuracy on 50 withheld sentences and generated NCSilver, a set of 98 newly translated modern Chinese sentences of varying lengths. A sample of both NCGold and NCSilver is included in the Supplementary Materials. Additionally, we developed FastText-based and Seq2Seq models to further support research on Nushu. NushuRescue provides a versatile and scalable tool for the revitalization of endangered languages, minimizing the need for extensive human input.
comment: Accepted to COLING 2025
♻ ☆ Tube Loss: A Novel Approach for Prediction Interval Estimation and probabilistic forecasting
This paper proposes a novel loss function, called 'Tube Loss', for simultaneous estimation of bounds of a Prediction Interval (PI) in the regression setup, and also for generating probabilistic forecasts from time series data solving a single optimization problem. The PIs obtained by minimizing the empirical risk based on the Tube Loss are shown to be of better quality than the PIs obtained by the existing methods in the following sense. First, it yields intervals that attain the prespecified confidence level $t \in(0,1)$ asymptotically. A theoretical proof of this fact is given. Secondly, the user is allowed to move the interval up or down by controlling the value of a parameter. This helps the user to choose a PI capturing denser regions of the probability distribution of the response variable inside the interval, and thus, sharpening its width. This is shown to be especially useful when the conditional distribution of the response variable is skewed. Further, the Tube Loss based PI estimation method can trade-off between the coverage and the average width by solving a single optimization problem. It enables further reduction of the average width of PI through re-calibration. Also, unlike a few existing PI estimation methods the gradient descent (GD) method can be used for minimization of empirical risk. Finally, through extensive experimentation, we have shown the efficacy of the Tube Loss based PI estimation in kernel machines, neural networks and deep networks and also for probabilistic forecasting tasks. The codes of the experiments are available at https://github.com/ltpritamanand/Tube_loss
♻ ☆ Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance
Recent controllable generation approaches such as FreeControl and Diffusion Self-Guidance bring fine-grained spatial and appearance control to text-to-image (T2I) diffusion models without training auxiliary modules. However, these methods optimize the latent embedding for each type of score function with longer diffusion steps, making the generation process time-consuming and limiting their flexibility and use. This work presents Ctrl-X, a simple framework for T2I diffusion controlling structure and appearance without additional training or guidance. Ctrl-X designs feed-forward structure control to enable the structure alignment with a structure image and semantic-aware appearance transfer to facilitate the appearance transfer from a user-input image. Extensive qualitative and quantitative experiments illustrate the superior performance of Ctrl-X on various condition inputs and model checkpoints. In particular, Ctrl-X supports novel structure and appearance control with arbitrary condition images of any modality, exhibits superior image quality and appearance transfer compared to existing works, and provides instant plug-and-play functionality to any T2I and text-to-video (T2V) diffusion model. See our project page for an overview of the results: https://genforce.github.io/ctrl-x
comment: 22 pages, 17 figures, see project page at https://genforce.github.io/ctrl-x
♻ ☆ LABOR-LLM: Language-Based Occupational Representations with Large Language Models
Vafa et al. (2024) introduced a transformer-based econometric model, CAREER, that predicts a worker's next job as a function of career history (an "occupation model"). CAREER was initially estimated ("pre-trained") using a large, unrepresentative resume dataset, which served as a "foundation model," and parameter estimation was continued ("fine-tuned") using data from a representative survey. CAREER had better predictive performance than benchmarks. This paper considers an alternative where the resume-based foundation model is replaced by a large language model (LLM). We convert tabular data from the survey into text files that resemble resumes and fine-tune the LLMs using these text files with the objective to predict the next token (word). The resulting fine-tuned LLM is used as an input to an occupation model. Its predictive performance surpasses all prior models. We demonstrate the value of fine-tuning and further show that by adding more career data from a different population, fine-tuning smaller LLMs surpasses the performance of fine-tuning larger models.
♻ ☆ ConceptSearch: Towards Efficient Program Search Using LLMs for Abstraction and Reasoning Corpus (ARC) AAAI 2025
The Abstraction and Reasoning Corpus (ARC) poses a significant challenge to artificial intelligence, demanding broad generalization and few-shot learning capabilities that remain elusive for current deep learning methods, including large language models (LLMs). While LLMs excel in program synthesis, their direct application to ARC yields limited success. To address this, we introduce ConceptSearch, a novel function-search algorithm that leverages LLMs for program generation and employs a concept-based scoring method to guide the search efficiently. Unlike simplistic pixel-based metrics like Hamming distance, ConceptSearch evaluates programs on their ability to capture the underlying transformation concept reflected in the input-output examples. We explore three scoring functions: Hamming distance, a CNN-based scoring function, and an LLM-based natural language scoring function. Experimental results demonstrate the effectiveness of ConceptSearch, achieving a significant performance improvement over direct prompting with GPT-4. Moreover, our novel concept-based scoring exhibits up to 30% greater efficiency compared to Hamming distance, measured in terms of the number of iterations required to reach the correct solution. These findings highlight the potential of LLM-driven program search when integrated with concept-based guidance for tackling challenging generalization problems like ARC.
comment: Pre-print of paper accepted at AAAI 2025
♻ ☆ Learnable Sparse Customization in Heterogeneous Edge Computing
To effectively manage and utilize massive distributed data at the network edge, Federated Learning (FL) has emerged as a promising edge computing paradigm across data silos. However, FL still faces two challenges: system heterogeneity (i.e., the diversity of hardware resources across edge devices) and statistical heterogeneity (i.e., non-IID data). Although sparsification can extract diverse submodels for diverse clients, most sparse FL works either simply assign submodels with artificially-given rigid rules or prune partial parameters using heuristic strategies, resulting in inflexible sparsification and poor performance. In this work, we propose Learnable Personalized Sparsification for heterogeneous Federated learning (FedLPS), which achieves the learnable customization of heterogeneous sparse models with importance-associated patterns and adaptive ratios to simultaneously tackle system and statistical heterogeneity. Specifically, FedLPS learns the importance of model units on local data representation and further derives an importance-based sparse pattern with minimal heuristics to accurately extract personalized data features in non-IID settings. Furthermore, Prompt Upper Confidence Bound Variance (P-UCBV) is designed to adaptively determine sparse ratios by learning the superimposed effect of diverse device capabilities and non-IID data, aiming at resource self-adaptation with promising accuracy. Extensive experiments show that FedLPS outperforms status quo approaches in accuracy and training costs, which improves accuracy by 1.28%-59.34% while reducing running time by more than 68.80%.
comment: There are some things to modify so we decided to withdraw first
♻ ☆ Optimal and Efficient Algorithms for Decentralized Online Convex Optimization
We investigate decentralized online convex optimization (D-OCO), in which a set of local learners are required to minimize a sequence of global loss functions using only local computations and communications. Previous studies have established $O(n^{5/4}\rho^{-1/2}\sqrt{T})$ and ${O}(n^{3/2}\rho^{-1}\log T)$ regret bounds for convex and strongly convex functions respectively, where $n$ is the number of local learners, $\rho<1$ is the spectral gap of the communication matrix, and $T$ is the time horizon. However, there exist large gaps from the existing lower bounds, i.e., $\Omega(n\sqrt{T})$ for convex functions and $\Omega(n)$ for strongly convex functions. To fill these gaps, in this paper, we first develop a novel D-OCO algorithm that can respectively reduce the regret bounds for convex and strongly convex functions to $\tilde{O}(n\rho^{-1/4}\sqrt{T})$ and $\tilde{O}(n\rho^{-1/2}\log T)$. The primary technique is to design an online accelerated gossip strategy that enjoys a faster average consensus among local learners. Furthermore, by carefully exploiting spectral properties of a specific network topology, we enhance the lower bounds for convex and strongly convex functions to $\Omega(n\rho^{-1/4}\sqrt{T})$ and $\Omega(n\rho^{-1/2}\log T)$, respectively. These results suggest that the regret of our algorithm is nearly optimal in terms of $T$, $n$, and $\rho$ for both convex and strongly convex functions. Finally, we propose a projection-free variant of our algorithm to efficiently handle practical applications with complex constraints. Our analysis reveals that the projection-free variant can achieve ${O}(nT^{3/4})$ and ${O}(nT^{2/3}(\log T)^{1/3})$ regret bounds for convex and strongly convex functions with nearly optimal $\tilde{O}(\rho^{-1/2}\sqrt{T})$ and $\tilde{O}(\rho^{-1/2}T^{1/3}(\log T)^{2/3})$ communication rounds, respectively.
♻ ☆ Exploring Consistency in Graph Representations:from Graph Kernels to Graph Neural Networks NeurIPS 2024
Graph Neural Networks (GNNs) have emerged as a dominant approach in graph representation learning, yet they often struggle to capture consistent similarity relationships among graphs. While graph kernel methods such as the Weisfeiler-Lehman subtree (WL-subtree) and Weisfeiler-Lehman optimal assignment (WLOA) kernels are effective in capturing similarity relationships, they rely heavily on predefined kernels and lack sufficient non-linearity for more complex data patterns. Our work aims to bridge the gap between neural network methods and kernel approaches by enabling GNNs to consistently capture relational structures in their learned representations. Given the analogy between the message-passing process of GNNs and WL algorithms, we thoroughly compare and analyze the properties of WL-subtree and WLOA kernels. We find that the similarities captured by WLOA at different iterations are asymptotically consistent, ensuring that similar graphs remain similar in subsequent iterations, thereby leading to superior performance over the WL-subtree kernel. Inspired by these findings, we conjecture that the consistency in the similarities of graph representations across GNN layers is crucial in capturing relational structures and enhancing graph classification performance. Thus, we propose a loss to enforce the similarity of graph representations to be consistent across different layers. Our empirical analysis verifies our conjecture and shows that our proposed consistency loss can significantly enhance graph classification performance across several GNN backbones on various datasets.
comment: NeurIPS 2024
♻ ☆ Dynamic Universal Approximation Theory: The Basic Theory for Transformer-based Large Language Models
Language models have emerged as a critical area of focus in artificial intelligence, particularly with the introduction of groundbreaking innovations like ChatGPT. Large-scale Transformer networks have quickly become the leading approach for advancing natural language processing algorithms. Built on the Transformer architecture, these models enable interactions that closely mimic human communication and, equipped with extensive knowledge, can even assist in guiding human tasks. Despite their impressive capabilities and growing complexity, a key question remains-the theoretical foundations of large language models (LLMs). What makes Transformer so effective for powering intelligent language applications, such as translation and coding? What underlies LLMs' ability for In-Context Learning (ICL)? How does the LoRA scheme enhance the fine-tuning of LLMs? And what supports the practicality of pruning LLMs? To address these critical questions and explore the technological strategies within LLMs, we leverage the Universal Approximation Theory (UAT) to offer a theoretical backdrop, shedding light on the mechanisms that underpin these advancements.
Artificial Intelligence 149
☆ GPD-1: Generative Pre-training for Driving
Modeling the evolutions of driving scenarios is important for the evaluation and decision-making of autonomous driving systems. Most existing methods focus on one aspect of scene evolution such as map generation, motion prediction, and trajectory planning. In this paper, we propose a unified Generative Pre-training for Driving (GPD-1) model to accomplish all these tasks altogether without additional fine-tuning. We represent each scene with ego, agent, and map tokens and formulate autonomous driving as a unified token generation problem. We adopt the autoregressive transformer architecture and use a scene-level attention mask to enable intra-scene bi-directional interactions. For the ego and agent tokens, we propose a hierarchical positional tokenizer to effectively encode both 2D positions and headings. For the map tokens, we train a map vector-quantized autoencoder to efficiently compress ego-centric semantic maps into discrete tokens. We pre-train our GPD-1 on the large-scale nuPlan dataset and conduct extensive experiments to evaluate its effectiveness. With different prompts, our GPD-1 successfully generalizes to various tasks without finetuning, including scene generation, traffic simulation, closed-loop simulation, map prediction, and motion planning. Code: https://github.com/wzzheng/GPD.
comment: Code is available at: https://github.com/wzzheng/GPD
☆ DMin: Scalable Training Data Influence Estimation for Diffusion Models
Identifying the training data samples that most influence a generated image is a critical task in understanding diffusion models, yet existing influence estimation methods are constrained to small-scale or LoRA-tuned models due to computational limitations. As diffusion models scale up, these methods become impractical. To address this challenge, we propose DMin (Diffusion Model influence), a scalable framework for estimating the influence of each training data sample on a given generated image. By leveraging efficient gradient compression and retrieval techniques, DMin reduces storage requirements from 339.39 TB to only 726 MB and retrieves the top-k most influential training samples in under 1 second, all while maintaining performance. Our empirical results demonstrate DMin is both effective in identifying influential training samples and efficient in terms of computational and storage requirements.
comment: 14 pages, 6 figures, 8 tables. Under Review
☆ Synthetic Vision: Training Vision-Language Models to Understand Physics
Physical reasoning, which involves the interpretation, understanding, and prediction of object behavior in dynamic environments, remains a significant challenge for current Vision-Language Models (VLMs). In this work, we propose two methods to enhance VLMs' physical reasoning capabilities using simulated data. First, we fine-tune a pre-trained VLM using question-answer (QA) pairs generated from simulations relevant to physical reasoning tasks. Second, we introduce Physics Context Builders (PCBs), specialized VLMs fine-tuned to create scene descriptions enriched with physical properties and processes. During physical reasoning tasks, these PCBs can be leveraged as context to assist a Large Language Model (LLM) to improve its performance. We evaluate both of our approaches using multiple benchmarks, including a new stability detection QA dataset called Falling Tower, which includes both simulated and real-world scenes, and CLEVRER. We demonstrate that a small QA fine-tuned VLM can significantly outperform larger state-of-the-art foundational models. We also show that integrating PCBs boosts the performance of foundational LLMs on physical reasoning tasks. Using the real-world scenes from the Falling Tower dataset, we also validate the robustness of both approaches in Sim2Real transfer. Our results highlight the utility that simulated data can have in the creation of learning systems capable of advanced physical reasoning.
☆ Image Retrieval Methods in the Dissimilarity Space
Image retrieval methods rely on metric learning to train backbone feature extraction models that can extract discriminant queries and reference (gallery) feature representations for similarity matching. Although state-of-the-art accuracy has improved considerably with the advent of deep learning (DL) models trained on large datasets, image retrieval remains challenging in many real-world video analytics and surveillance applications, e.g., person re-identification. Using the Euclidean space for matching limits the performance in real-world applications due to the curse of dimensionality, overfitting, and sensitivity to noisy data. We argue that the feature dissimilarity space is more suitable for similarity matching, and propose a dichotomy transformation to project query and reference embeddings into a single embedding in the dissimilarity space. We also advocate for end-to-end training of a backbone and binary classification models for pair-wise matching. As opposed to comparing the distance between queries and reference embeddings, we show the benefits of classifying the single dissimilarity space embedding (as similar or dissimilar), especially when trained end-to-end. We propose a method to train the max-margin classifier together with the backbone feature extractor by applying constraints to the L2 norm of the classifier weights along with the hinge loss. Our extensive experiments on challenging image retrieval datasets and using diverse feature extraction backbones highlight the benefits of similarity matching in the dissimilarity space. In particular, when jointly training the feature extraction backbone and regularised classifier for matching, the dissimilarity space provides a higher level of accuracy.
comment: 7 pages
☆ Competition and Diversity in Generative AI
Recent evidence suggests that the use of generative artificial intelligence reduces the diversity of content produced. In this work, we develop a game-theoretic model to explore the downstream consequences of content homogeneity when producers use generative AI to compete with one another. At equilibrium, players indeed produce content that is less diverse than optimal. However, stronger competition mitigates homogeneity and induces more diverse production. Perhaps more surprisingly, we show that a generative AI model that performs well in isolation (i.e., according to a benchmark) may fail to do so when faced with competition, and vice versa. We validate our results empirically by using language models to play Scattergories, a word game in which players are rewarded for producing answers that are both correct and unique. We discuss how the interplay between competition and homogeneity has implications for the development, evaluation, and use of generative AI.
☆ AdvWave: Stealthy Adversarial Jailbreak Attack against Large Audio-Language Models
Recent advancements in large audio-language models (LALMs) have enabled speech-based user interactions, significantly enhancing user experience and accelerating the deployment of LALMs in real-world applications. However, ensuring the safety of LALMs is crucial to prevent risky outputs that may raise societal concerns or violate AI regulations. Despite the importance of this issue, research on jailbreaking LALMs remains limited due to their recent emergence and the additional technical challenges they present compared to attacks on DNN-based audio models. Specifically, the audio encoders in LALMs, which involve discretization operations, often lead to gradient shattering, hindering the effectiveness of attacks relying on gradient-based optimizations. The behavioral variability of LALMs further complicates the identification of effective (adversarial) optimization targets. Moreover, enforcing stealthiness constraints on adversarial audio waveforms introduces a reduced, non-convex feasible solution space, further intensifying the challenges of the optimization process. To overcome these challenges, we develop AdvWave, the first jailbreak framework against LALMs. We propose a dual-phase optimization method that addresses gradient shattering, enabling effective end-to-end gradient-based optimization. Additionally, we develop an adaptive adversarial target search algorithm that dynamically adjusts the adversarial optimization target based on the response patterns of LALMs for specific queries. To ensure that adversarial audio remains perceptually natural to human listeners, we design a classifier-guided optimization approach that generates adversarial noise resembling common urban sounds. Extensive evaluations on multiple advanced LALMs demonstrate that AdvWave outperforms baseline methods, achieving a 40% higher average jailbreak attack success rate.
☆ Preference Discerning with LLM-Enhanced Generative Retrieval
Sequential recommendation systems aim to provide personalized recommendations for users based on their interaction history. To achieve this, they often incorporate auxiliary information, such as textual descriptions of items and auxiliary tasks, like predicting user preferences and intent. Despite numerous efforts to enhance these models, they still suffer from limited personalization. To address this issue, we propose a new paradigm, which we term preference discerning. In preference dscerning, we explicitly condition a generative sequential recommendation system on user preferences within its context. To this end, we generate user preferences using Large Language Models (LLMs) based on user reviews and item-specific data. To evaluate preference discerning capabilities of sequential recommendation systems, we introduce a novel benchmark that provides a holistic evaluation across various scenarios, including preference steering and sentiment following. We assess current state-of-the-art methods using our benchmark and show that they struggle to accurately discern user preferences. Therefore, we propose a new method named Mender ($\textbf{M}$ultimodal Prefer$\textbf{en}$ce $\textbf{d}$iscern$\textbf{er}$), which improves upon existing methods and achieves state-of-the-art performance on our benchmark. Our results show that Mender can be effectively guided by human preferences even though they have not been observed during training, paving the way toward more personalized sequential recommendation systems. We will open-source the code and benchmarks upon publication.
comment: 11 pages + references and appendix
☆ RoomTour3D: Geometry-Aware Video-Instruction Tuning for Embodied Navigation
Vision-and-Language Navigation (VLN) suffers from the limited diversity and scale of training data, primarily constrained by the manual curation of existing simulators. To address this, we introduce RoomTour3D, a video-instruction dataset derived from web-based room tour videos that capture real-world indoor spaces and human walking demonstrations. Unlike existing VLN datasets, RoomTour3D leverages the scale and diversity of online videos to generate open-ended human walking trajectories and open-world navigable instructions. To compensate for the lack of navigation data in online videos, we perform 3D reconstruction and obtain 3D trajectories of walking paths augmented with additional information on the room types, object locations and 3D shape of surrounding scenes. Our dataset includes $\sim$100K open-ended description-enriched trajectories with $\sim$200K instructions, and 17K action-enriched trajectories from 1847 room tour environments. We demonstrate experimentally that RoomTour3D enables significant improvements across multiple VLN tasks including CVDN, SOON, R2R, and REVERIE. Moreover, RoomTour3D facilitates the development of trainable zero-shot VLN agents, showcasing the potential and challenges of advancing towards open-world navigation.
☆ Advancing Single- and Multi-task Text Classification through Large Language Model Fine-tuning
Both encoder-only models (e.g., BERT, RoBERTa) and large language models (LLMs, e.g., Llama3) have been widely used for text classification tasks. However, there is a lack of systematic studies comparing the performance of encoder-based models and LLMs in text classification, particularly when fine-tuning is involved. This study employed a diverse range of models and methods, varying in size and architecture, and including both fine-tuned and pre-trained approaches. We first assessed the performances of these LLMs on the 20 Newsgroups (20NG) and MASSIVE datasets, comparing them to encoder-only RoBERTa models. Additionally, we explored the multi-task capabilities of both model types by combining multiple classification tasks, including intent detection and slot-filling, into a single model using data from both datasets. Our results indicate that fully fine-tuned Llama3-70B models outperform RoBERTa-large and other decoder LLMs across various classification tasks and datasets. Moreover, the consolidated multi-task fine-tuned LLMs matched the performance of dual-model setups in both tasks across both datasets. Overall, our study provides a comprehensive benchmark of encoder-only and LLM models on text classification tasks and demonstrates a method to combine two or more fully fine-tuned decoder LLMs for reduced latency and equivalent performance.
comment: 9 pages, 3 tables
☆ TURBOATTENTION: Efficient Attention Approximation For High Throughputs LLMs
Large language model (LLM) inference demands significant amount of computation and memory, especially in the key attention mechanism. While techniques, such as quantization and acceleration algorithms, like FlashAttention, have improved efficiency of the overall inference, they address different aspects of the problem: quantization focuses on weight-activation operations, while FlashAttention improves execution but requires high-precision formats. Recent Key-value (KV) cache quantization reduces memory bandwidth but still needs floating-point dequantization for attention operation. We present TurboAttention, a comprehensive approach to enable quantized execution of attention that simultaneously addresses both memory and computational efficiency. Our solution introduces two key innovations: FlashQ, a headwise attention quantization technique that enables both compression of KV cache and quantized execution of activation-activation multiplication, and Sparsity-based Softmax Approximation (SAS), which eliminates the need for dequantization to FP32 during exponentiation operation in attention. Experimental results demonstrate that TurboAttention achieves 1.2-1.8x speedup in attention, reduces the KV cache size by over 4.4x, and enables up to 2.37x maximum throughput over the FP16 baseline while outperforming state-of-the-art quantization and compression techniques across various datasets and models.
☆ Learning Sketch Decompositions in Planning via Deep Reinforcement Learning
In planning and reinforcement learning, the identification of common subgoal structures across problems is important when goals are to be achieved over long horizons. Recently, it has been shown that such structures can be expressed as feature-based rules, called sketches, over a number of classical planning domains. These sketches split problems into subproblems which then become solvable in low polynomial time by a greedy sequence of IW$(k)$ searches. Methods for learning sketches using feature pools and min-SAT solvers have been developed, yet they face two key limitations: scalability and expressivity. In this work, we address these limitations by formulating the problem of learning sketch decompositions as a deep reinforcement learning (DRL) task, where general policies are sought in a modified planning problem where the successor states of a state s are defined as those reachable from s through an IW$(k)$ search. The sketch decompositions obtained through this method are experimentally evaluated across various domains, and problems are regarded as solved by the decomposition when the goal is reached through a greedy sequence of IW$(k)$ searches. While our DRL approach for learning sketch decompositions does not yield interpretable sketches in the form of rules, we demonstrate that the resulting decompositions can often be understood in a crisp manner.
☆ GenPlan: Generative sequence models as adaptive planners AAAI 2025
Offline reinforcement learning has shown tremendous success in behavioral planning by learning from previously collected demonstrations. However, decision-making in multitask missions still presents significant challenges. For instance, a mission might require an agent to explore an unknown environment, discover goals, and navigate to them, even if it involves interacting with obstacles along the way. Such behavioral planning problems are difficult to solve due to: a) agents failing to adapt beyond the single task learned through their reward function, and b) the inability to generalize to new environments not covered in the training demonstrations, e.g., environments where all doors were unlocked in the demonstrations. Consequently, state-of-the-art decision making methods are limited to missions where the required tasks are well-represented in the training demonstrations and can be solved within a short (temporal) planning horizon. To address this, we propose GenPlan: a stochastic and adaptive planner that leverages discrete-flow models for generative sequence modeling, enabling sample-efficient exploration and exploitation. This framework relies on an iterative denoising procedure to generate a sequence of goals and actions. This approach captures multi-modal action distributions and facilitates goal and task discovery, thereby enhancing generalization to out-of-distribution tasks and environments, i.e., missions not part of the training data. We demonstrate the effectiveness of our method through multiple simulation environments. Notably, GenPlan outperforms the state-of-the-art methods by over 10% on adaptive planning tasks, where the agent adapts to multi-task missions while leveraging demonstrations on single-goal-reaching tasks.
comment: Accepted in AAAI 2025. Project page: https://aku02.github.io/projects/genplan/
☆ Exact Algorithms for Multiagent Path Finding with Communication Constraints on Tree-Like Structures
Consider the scenario where multiple agents have to move in an optimal way through a network, each one towards their ending position while avoiding collisions. By optimal, we mean as fast as possible, which is evaluated by a measure known as the makespan of the proposed solution. This is the setting studied in the Multiagent Path Finding problem. In this work, we additionally provide the agents with a way to communicate with each other. Due to size constraints, it is reasonable to assume that the range of communication of each agent will be limited. What should be the trajectories of the agents to, additionally, maintain a backbone of communication? In this work, we study the Multiagent Path Finding with Communication Constraint problem under the parameterized complexity framework. Our main contribution is three exact algorithms that are efficient when considering particular structures for the input network. We provide such algorithms for the case when the communication range and the number of agents (the makespan resp.) are provided in the input and the network has a tree topology, or bounded maximum degree (has a tree-like topology, i.e., bounded treewidth resp.). We complement these results by showing that it is highly unlikely to construct efficient algorithms when considering the number of agents as part of the input, even if the makespan is $3$ and the communication range is $1$.
☆ MaestroMotif: Skill Design from Artificial Intelligence Feedback
Describing skills in natural language has the potential to provide an accessible way to inject human knowledge about decision-making into an AI system. We present MaestroMotif, a method for AI-assisted skill design, which yields high-performing and adaptable agents. MaestroMotif leverages the capabilities of Large Language Models (LLMs) to effectively create and reuse skills. It first uses an LLM's feedback to automatically design rewards corresponding to each skill, starting from their natural language description. Then, it employs an LLM's code generation abilities, together with reinforcement learning, for training the skills and combining them to implement complex behaviors specified in language. We evaluate MaestroMotif using a suite of complex tasks in the NetHack Learning Environment (NLE), demonstrating that it surpasses existing approaches in both performance and usability.
☆ GR-NLP-TOOLKIT: An Open-Source NLP Toolkit for Modern Greek COLING 2025
We present GR-NLP-TOOLKIT, an open-source natural language processing (NLP) toolkit developed specifically for modern Greek. The toolkit provides state-of-the-art performance in five core NLP tasks, namely part-of-speech tagging, morphological tagging, dependency parsing, named entity recognition, and Greeklishto-Greek transliteration. The toolkit is based on pre-trained Transformers, it is freely available, and can be easily installed in Python (pip install gr-nlp-toolkit). It is also accessible through a demonstration platform on HuggingFace, along with a publicly available API for non-commercial use. We discuss the functionality provided for each task, the underlying methods, experiments against comparable open-source toolkits, and future possible enhancements. The toolkit is available at: https://github.com/nlpaueb/gr-nlp-toolkit
comment: Accepted Demo Paper @ COLING 2025 (Github: https://github.com/nlpaueb/gr-nlp-toolkit/, Demo: https://huggingface.co/spaces/AUEB-NLP/greek-nlp-toolkit-demo, API: https://huggingface.co/spaces/AUEB-NLP/The-Greek-NLP-API)
☆ Enhancing Interpretability Through Loss-Defined Classification Objective in Structured Latent Spaces
Supervised machine learning often operates on the data-driven paradigm, wherein internal model parameters are autonomously optimized to converge predicted outputs with the ground truth, devoid of explicitly programming rules or a priori assumptions. Although data-driven methods have yielded notable successes across various benchmark datasets, they inherently treat models as opaque entities, thereby limiting their interpretability and yielding a lack of explanatory insights into their decision-making processes. In this work, we introduce Latent Boost, a novel approach that integrates advanced distance metric learning into supervised classification tasks, enhancing both interpretability and training efficiency. Thus during training, the model is not only optimized for classification metrics of the discrete data points but also adheres to the rule that the collective representation zones of each class should be sharply clustered. By leveraging the rich structural insights of intermediate model layer latent representations, Latent Boost improves classification interpretability, as demonstrated by higher Silhouette scores, while accelerating training convergence. These performance and latent structural benefits are achieved with minimum additional cost, making it broadly applicable across various datasets without requiring data-specific adjustments. Furthermore, Latent Boost introduces a new paradigm for aligning classification performance with improved model transparency to address the challenges of black-box models.
☆ REPEAT: Improving Uncertainty Estimation in Representation Learning Explainability AAAI 2025
Incorporating uncertainty is crucial to provide trustworthy explanations of deep learning models. Recent works have demonstrated how uncertainty modeling can be particularly important in the unsupervised field of representation learning explainable artificial intelligence (R-XAI). Current R-XAI methods provide uncertainty by measuring variability in the importance score. However, they fail to provide meaningful estimates of whether a pixel is certainly important or not. In this work, we propose a new R-XAI method called REPEAT that addresses the key question of whether or not a pixel is \textit{certainly} important. REPEAT leverages the stochasticity of current R-XAI methods to produce multiple estimates of importance, thus considering each pixel in an image as a Bernoulli random variable that is either important or unimportant. From these Bernoulli random variables we can directly estimate the importance of a pixel and its associated certainty, thus enabling users to determine certainty in pixel importance. Our extensive evaluation shows that REPEAT gives certainty estimates that are more intuitive, better at detecting out-of-distribution data, and more concise.
comment: Accepted at AAAI 2025. Code available at: https://github.com/Wickstrom/REPEAT
☆ PointTalk: Audio-Driven Dynamic Lip Point Cloud for 3D Gaussian-based Talking Head Synthesis AAAI 2025
Talking head synthesis with arbitrary speech audio is a crucial challenge in the field of digital humans. Recently, methods based on radiance fields have received increasing attention due to their ability to synthesize high-fidelity and identity-consistent talking heads from just a few minutes of training video. However, due to the limited scale of the training data, these methods often exhibit poor performance in audio-lip synchronization and visual quality. In this paper, we propose a novel 3D Gaussian-based method called PointTalk, which constructs a static 3D Gaussian field of the head and deforms it in sync with the audio. It also incorporates an audio-driven dynamic lip point cloud as a critical component of the conditional information, thereby facilitating the effective synthesis of talking heads. Specifically, the initial step involves generating the corresponding lip point cloud from the audio signal and capturing its topological structure. The design of the dynamic difference encoder aims to capture the subtle nuances inherent in dynamic lip movements more effectively. Furthermore, we integrate the audio-point enhancement module, which not only ensures the synchronization of the audio signal with the corresponding lip point cloud within the feature space, but also facilitates a deeper understanding of the interrelations among cross-modal conditional features. Extensive experiments demonstrate that our method achieves superior high-fidelity and audio-lip synchronization in talking head synthesis compared to previous methods.
comment: 9 pages, accepted by AAAI 2025
☆ SuperCode: Sustainability PER AI-driven CO-DEsign
Currently, data-intensive scientific applications require vast amounts of compute resources to deliver world-leading science. The climate emergency has made it clear that unlimited use of resources (e.g., energy) for scientific discovery is no longer acceptable. Future computing hardware promises to be much more energy efficient, but without better optimized software this cannot reach its full potential. In this vision paper, we propose a generic AI-driven co-design methodology, using specialized Large Language Models (like ChatGPT), to effectively generate efficient code for emerging computing hardware. We describe how we will validate our methodology with two radio astronomy applications, with sustainability as the key performance indicator. This paper is a modified version of our accepted SuperCode project proposal. We present it here in this form to introduce the vision behind this project and to disseminate the work in the spirit of Open Science and transparency. An additional aim is to collect feedback, invite potential collaboration partners and use-cases to join the project.
☆ Accurate Water Level Monitoring in AWD Rice Cultivation Using Convolutional Neural Networks
The Alternate Wetting and Drying (AWD) method is a rice-growing water management technique promoted as a sustainable alternative to Continuous Flooding (CF). Climate change has placed the agricultural sector in a challenging position, particularly as global water resources become increasingly scarce, affecting rice production on irrigated lowlands. Rice, a staple food for over half of the world's population, demands significantly more water than other major crops. In Bangladesh, \textit{Boro} rice, in particular, requires considerable water inputs during its cultivation. Traditionally, farmers manually measure water levels, a process that is both time-consuming and prone to errors. While ultrasonic sensors offer improvements in water height measurement, they still face limitations, such as susceptibility to weather conditions and environmental factors. To address these issues, we propose a novel approach that automates water height measurement using computer vision, specifically through a convolutional neural network (CNN). Our attention-based architecture achieved an $R^2$ score of 0.9885 and a Mean Squared Error (MSE) of 0.2766, providing a more accurate and efficient solution for managing AWD systems.
comment: 15 pages, 6 figures
☆ Bootstrapping Language-Guided Navigation Learning with Self-Refining Data Flywheel
Creating high-quality data for training robust language-instructed agents is a long-lasting challenge in embodied AI. In this paper, we introduce a Self-Refining Data Flywheel (SRDF) that generates high-quality and large-scale navigational instruction-trajectory pairs by iteratively refining the data pool through the collaboration between two models, the instruction generator and the navigator, without any human-in-the-loop annotation. Specifically, SRDF starts with using a base generator to create an initial data pool for training a base navigator, followed by applying the trained navigator to filter the data pool. This leads to higher-fidelity data to train a better generator, which can, in turn, produce higher-quality data for training the next-round navigator. Such a flywheel establishes a data self-refining process, yielding a continuously improved and highly effective dataset for large-scale language-guided navigation learning. Our experiments demonstrate that after several flywheel rounds, the navigator elevates the performance boundary from 70% to 78% SPL on the classic R2R test set, surpassing human performance (76%) for the first time. Meanwhile, this process results in a superior generator, evidenced by a SPICE increase from 23.5 to 26.2, better than all previous VLN instruction generation methods. Finally, we demonstrate the scalability of our method through increasing environment and instruction diversity, and the generalization ability of our pre-trained navigator across various downstream navigation tasks, surpassing state-of-the-art methods by a large margin in all cases.
comment: 28 pages, Code and data are available at https://github.com/wz0919/VLN-SRDF
☆ IRL for Restless Multi-Armed Bandits with Applications in Maternal and Child Health
Public health practitioners often have the goal of monitoring patients and maximizing patients' time spent in "favorable" or healthy states while being constrained to using limited resources. Restless multi-armed bandits (RMAB) are an effective model to solve this problem as they are helpful to allocate limited resources among many agents under resource constraints, where patients behave differently depending on whether they are intervened on or not. However, RMABs assume the reward function is known. This is unrealistic in many public health settings because patients face unique challenges and it is impossible for a human to know who is most deserving of any intervention at such a large scale. To address this shortcoming, this paper is the first to present the use of inverse reinforcement learning (IRL) to learn desired rewards for RMABs, and we demonstrate improved outcomes in a maternal and child health telehealth program. First we allow public health experts to specify their goals at an aggregate or population level and propose an algorithm to design expert trajectories at scale based on those goals. Second, our algorithm WHIRL uses gradient updates to optimize the objective, allowing for efficient and accurate learning of RMAB rewards. Third, we compare with existing baselines and outperform those in terms of run-time and accuracy. Finally, we evaluate and show the usefulness of WHIRL on thousands on beneficiaries from a real-world maternal and child health setting in India. We publicly release our code here: https://github.com/Gjain234/WHIRL.
☆ Federated Learning for Traffic Flow Prediction with Synthetic Data Augmentation
Deep-learning based traffic prediction models require vast amounts of data to learn embedded spatial and temporal dependencies. The inherent privacy and commercial sensitivity of such data has encouraged a shift towards decentralised data-driven methods, such as Federated Learning (FL). Under a traditional Machine Learning paradigm, traffic flow prediction models can capture spatial and temporal relationships within centralised data. In reality, traffic data is likely distributed across separate data silos owned by multiple stakeholders. In this work, a cross-silo FL setting is motivated to facilitate stakeholder collaboration for optimal traffic flow prediction applications. This work introduces an FL framework, referred to as FedTPS, to generate synthetic data to augment each client's local dataset by training a diffusion-based trajectory generation model through FL. The proposed framework is evaluated on a large-scale real world ride-sharing dataset using various FL methods and Traffic Flow Prediction models, including a novel prediction model we introduce, which leverages Temporal and Graph Attention mechanisms to learn the Spatio-Temporal dependencies embedded within regional traffic flow data. Experimental results show that FedTPS outperforms multiple other FL baselines with respect to global model performance.
comment: 11 pages, 7 figures, 6 tables, ACM format
☆ Efficient Rectification of Neuro-Symbolic Reasoning Inconsistencies by Abductive Reflection AAAI 2025
Neuro-Symbolic (NeSy) AI could be regarded as an analogy to human dual-process cognition, modeling the intuitive System 1 with neural networks and the algorithmic System 2 with symbolic reasoning. However, for complex learning targets, NeSy systems often generate outputs inconsistent with domain knowledge and it is challenging to rectify them. Inspired by the human Cognitive Reflection, which promptly detects errors in our intuitive response and revises them by invoking the System 2 reasoning, we propose to improve NeSy systems by introducing Abductive Reflection (ABL-Refl) based on the Abductive Learning (ABL) framework. ABL-Refl leverages domain knowledge to abduce a reflection vector during training, which can then flag potential errors in the neural network outputs and invoke abduction to rectify them and generate consistent outputs during inference. ABL-Refl is highly efficient in contrast to previous ABL implementations. Experiments show that ABL-Refl outperforms state-of-the-art NeSy methods, achieving excellent accuracy with fewer training resources and enhanced efficiency.
comment: Accepted to AAAI 2025
☆ TapeAgents: a Holistic Framework for Agent Development and Optimization
We present TapeAgents, an agent framework built around a granular, structured log tape of the agent session that also plays the role of the session's resumable state. In TapeAgents we leverage tapes to facilitate all stages of the LLM Agent development lifecycle. The agent reasons by processing the tape and the LLM output to produce new thought and action steps and append them to the tape. The environment then reacts to the agent's actions by likewise appending observation steps to the tape. By virtue of this tape-centred design, TapeAgents can provide AI practitioners with holistic end-to-end support. At the development stage, tapes facilitate session persistence, agent auditing, and step-by-step debugging. Post-deployment, one can reuse tapes for evaluation, fine-tuning, and prompt-tuning; crucially, one can adapt tapes from other agents or use revised historical tapes. In this report, we explain the TapeAgents design in detail. We demonstrate possible applications of TapeAgents with several concrete examples of building monolithic agents and multi-agent teams, of optimizing agent prompts and finetuning the agent's LLM. We present tooling prototypes and report a case study where we use TapeAgents to finetune a Llama-3.1-8B form-filling assistant to perform as well as GPT-4o while being orders of magnitude cheaper. Lastly, our comparative analysis shows that TapeAgents's advantages over prior frameworks stem from our novel design of the LLM agent as a resumable, modular state machine with a structured configuration, that generates granular, structured logs and that can transform these logs into training text -- a unique combination of features absent in previous work.
☆ Proactive Model Adaptation Against Concept Drift for Online Time Series Forecasting KDD 2025
Time series forecasting always faces the challenge of concept drift, where data distributions evolve over time, leading to a decline in forecast model performance. Existing solutions are based on online learning, which continually organize recent time series observations as new training samples and update model parameters according to the forecasting feedback on recent data. However, they overlook a critical issue: obtaining ground-truth future values of each sample should be delayed until after the forecast horizon. This delay creates a temporal gap between the training samples and the test sample. Our empirical analysis reveals that the gap can introduce concept drift, causing forecast models to adapt to outdated concepts. In this paper, we present \textsc{Proceed}, a novel proactive model adaptation framework for online time series forecasting. \textsc{Proceed} first operates by estimating the concept drift between the recently used training samples and the current test sample. It then employs an adaptation generator to efficiently translate the estimated drift into parameter adjustments, proactively adapting the model to the test sample. To enhance the generalization capability of the framework, \textsc{Proceed} is trained on synthetic diverse concept drifts. We conduct extensive experiments on five real-world datasets across various forecast models. The empirical study demonstrates that our proposed \textsc{Proceed} brings more performance improvements than the state-of-the-art online learning methods, significantly facilitating forecast models' resilience against concept drifts.
comment: Accepted by KDD 2025. Preprint version
☆ Mitigating Out-of-Entity Errors in Named Entity Recognition: A Sentence-Level Strategy COLING 2025
Many previous models of named entity recognition (NER) suffer from the problem of Out-of-Entity (OOE), i.e., the tokens in the entity mentions of the test samples have not appeared in the training samples, which hinders the achievement of satisfactory performance. To improve OOE-NER performance, in this paper, we propose a new framework, namely S+NER, which fully leverages sentence-level information. Our S+NER achieves better OOE-NER performance mainly due to the following two particular designs. 1) It first exploits the pre-trained language model's capability of understanding the target entity's sentence-level context with a template set. 2) Then, it refines the sentence-level representation based on the positive and negative templates, through a contrastive learning strategy and template pooling method, to obtain better NER results. Our extensive experiments on five benchmark datasets have demonstrated that, our S+NER outperforms some state-of-the-art OOE-NER models.
comment: Accepted by COLING 2025
☆ SwarmGPT-Primitive: A Language-Driven Choreographer for Drone Swarms Using Safe Motion Primitive Composition ICRA 2025
Catalyzed by advancements in hardware and software, drone performances are increasingly making their mark in the entertainment industry. However, designing smooth and safe choreographies for drone swarms is complex and often requires expert domain knowledge. In this work, we introduce SwarmGPT-Primitive, a language-based choreographer that integrates the reasoning capabilities of large language models (LLMs) with safe motion planning to facilitate deployable drone swarm choreographies. The LLM composes choreographies for a given piece of music by utilizing a library of motion primitives; the language-based choreographer is augmented with an optimization-based safety filter, which certifies the choreography for real-world deployment by making minimal adjustments when feasibility and safety constraints are violated. The overall SwarmGPT-Primitive framework decouples choreographic design from safe motion planning, which allows non-expert users to re-prompt and refine compositions without concerns about compliance with constraints such as avoiding collisions or downwash effects or satisfying actuation limits. We demonstrate our approach through simulations and experiments with swarms of up to 20 drones performing choreographies designed based on various songs, highlighting the system's ability to generate effective and synchronized drone choreographies for real-world deployment.
comment: Submitted to ICRA 2025
☆ The Roles of English in Evaluating Multilingual Language Models
Multilingual natural language processing is getting increased attention, with numerous models, benchmarks, and methods being released for many languages. English is often used in multilingual evaluation to prompt language models (LMs), mainly to overcome the lack of instruction tuning data in other languages. In this position paper, we lay out two roles of English in multilingual LM evaluations: as an interface and as a natural language. We argue that these roles have different goals: task performance versus language understanding. This discrepancy is highlighted with examples from datasets and evaluation setups. Numerous works explicitly use English as an interface to boost task performance. We recommend to move away from this imprecise method and instead focus on furthering language understanding.
comment: NoDaLiDa 2025
☆ NyayaAnumana & INLegalLlama: The Largest Indian Legal Judgment Prediction Dataset and Specialized Language Model for Enhanced Decision Analysis COLING 2025
The integration of artificial intelligence (AI) in legal judgment prediction (LJP) has the potential to transform the legal landscape, particularly in jurisdictions like India, where a significant backlog of cases burdens the legal system. This paper introduces NyayaAnumana, the largest and most diverse corpus of Indian legal cases compiled for LJP, encompassing a total of 7,02,945 preprocessed cases. NyayaAnumana, which combines the words "Nyay" (judgment) and "Anuman" (prediction or inference) respectively for most major Indian languages, includes a wide range of cases from the Supreme Court, High Courts, Tribunal Courts, District Courts, and Daily Orders and, thus, provides unparalleled diversity and coverage. Our dataset surpasses existing datasets like PredEx and ILDC, offering a comprehensive foundation for advanced AI research in the legal domain. In addition to the dataset, we present INLegalLlama, a domain-specific generative large language model (LLM) tailored to the intricacies of the Indian legal system. It is developed through a two-phase training approach over a base LLaMa model. First, Indian legal documents are injected using continual pretraining. Second, task-specific supervised finetuning is done. This method allows the model to achieve a deeper understanding of legal contexts. Our experiments demonstrate that incorporating diverse court data significantly boosts model accuracy, achieving approximately 90% F1-score in prediction tasks. INLegalLlama not only improves prediction accuracy but also offers comprehensible explanations, addressing the need for explainability in AI-assisted legal decisions.
comment: Accepted on COLING 2025
☆ HyViLM: Enhancing Fine-Grained Recognition with a Hybrid Encoder for Vision-Language Models
Recently, there has been growing interest in the capability of multimodal large language models (MLLMs) to process high-resolution images. A common approach currently involves dynamically cropping the original high-resolution image into smaller sub-images, which are then fed into a vision encoder that was pre-trained on lower-resolution images. However, this cropping approach often truncates objects and connected areas in the original image, causing semantic breaks. To address this limitation, we introduce HyViLM, designed to process images of any resolution while retaining the overall context during encoding. Specifically, we: (i) Design a new visual encoder called Hybrid Encoder that not only encodes individual sub-images but also interacts with detailed global visual features, significantly improving the model's ability to encode high-resolution images. (ii) Propose an optimal feature fusion strategy for the dynamic cropping approach, effectively leveraging information from different layers of the vision encoder. Compared with the state-of-the-art MLLMs under the same setting, our HyViLM outperforms existing MLLMs in nine out of ten tasks. Specifically, HyViLM achieves a 9.6% improvement in performance on the TextVQA task and a 6.9% enhancement on the DocVQA task.
comment: 11 pages, 4 figures
☆ Agency and Morality as part of Text Entry AI Assistant Personas
This paper discusses the need to move away from an instrumental view of text composition AI assistants under direct control of the user, towards a more agentic approach that is based on a value rationale. Based on an analysis of moral dimensions of AI assistance in computer mediated communication, the paper proposes basic guidelines for designing the agent's persona.
☆ SmolTulu: Higher Learning Rate to Batch Size Ratios Can Lead to Better Reasoning in SLMs
We present SmolTulu-1.7b-Instruct, referenced in this report as SmolTulu-DPO-1130, an instruction-tuned language model that adapts AllenAI's Tulu 3 post-training pipeline to enhance Huggingface's SmolLM2-1.7B base model. Through comprehensive empirical analysis using a 135M parameter model, we demonstrate that the relationship between learning rate and batch size significantly impacts model performance in a task-dependent manner. Our findings reveal a clear split: reasoning tasks like ARC and GSM8K benefit from higher learning rate to batch size ratios, while pattern recognition tasks such as HellaSwag and IFEval show optimal performance with lower ratios. These insights informed the development of SmolTulu, which achieves state-of-the-art performance among sub-2B parameter models on instruction following, scoring 67.7% on IFEval ($\Delta$11%), and mathematical reasoning with 51.6% on GSM8K ($\Delta$3.4%), with an alternate version achieving scoring 57.1% on ARC ($\Delta5.4%$). We release our model, training recipes, and ablation studies to facilitate further research in efficient model alignment, demonstrating that careful adaptation of optimization dynamics can help bridge the capability gap between small and large language models.
comment: 10 pages, 4 figures, and 13 tables. For the SmolTulu-1.7b-instruct model, see: https://huggingface.co/SultanR/SmolTulu-1.7b-Instruct
☆ Self-Refining Diffusion Samplers: Enabling Parallelization via Parareal Iterations NeurIPS 2024
In diffusion models, samples are generated through an iterative refinement process, requiring hundreds of sequential model evaluations. Several recent methods have introduced approximations (fewer discretization steps or distillation) to trade off speed at the cost of sample quality. In contrast, we introduce Self-Refining Diffusion Samplers (SRDS) that retain sample quality and can improve latency at the cost of additional parallel compute. We take inspiration from the Parareal algorithm, a popular numerical method for parallel-in-time integration of differential equations. In SRDS, a quick but rough estimate of a sample is first created and then iteratively refined in parallel through Parareal iterations. SRDS is not only guaranteed to accurately solve the ODE and converge to the serial solution but also benefits from parallelization across the diffusion trajectory, enabling batched inference and pipelining. As we demonstrate for pre-trained diffusion models, the early convergence of this refinement procedure drastically reduces the number of steps required to produce a sample, speeding up generation for instance by up to 1.7x on a 25-step StableDiffusion-v2 benchmark and up to 4.3x on longer trajectories.
comment: NeurIPS 2024
☆ How Does the Smoothness Approximation Method Facilitate Generalization for Federated Adversarial Learning?
Federated Adversarial Learning (FAL) is a robust framework for resisting adversarial attacks on federated learning. Although some FAL studies have developed efficient algorithms, they primarily focus on convergence performance and overlook generalization. Generalization is crucial for evaluating algorithm performance on unseen data. However, generalization analysis is more challenging due to non-smooth adversarial loss functions. A common approach to addressing this issue is to leverage smoothness approximation. In this paper, we develop algorithm stability measures to evaluate the generalization performance of two popular FAL algorithms: \textit{Vanilla FAL (VFAL)} and {\it Slack FAL (SFAL)}, using three different smooth approximation methods: 1) \textit{Surrogate Smoothness Approximation (SSA)}, (2) \textit{Randomized Smoothness Approximation (RSA)}, and (3) \textit{Over-Parameterized Smoothness Approximation (OPSA)}. Based on our in-depth analysis, we answer the question of how to properly set the smoothness approximation method to mitigate generalization error in FAL. Moreover, we identify RSA as the most effective method for reducing generalization error. In highly data-heterogeneous scenarios, we also recommend employing SFAL to mitigate the deterioration of generalization performance caused by heterogeneity. Based on our theoretical results, we provide insights to help develop more efficient FAL algorithms, such as designing new metrics and dynamic aggregation rules to mitigate heterogeneity.
☆ Can transformative AI shape a new age for our civilization?: Navigating between speculation and reality
Artificial Intelligence is widely regarded as a transformative force with the potential to redefine numerous sectors of human civilization. While Artificial Intelligence has evolved from speculative fiction to a pivotal element of technological progress, its role as a truly transformative agent, or transformative Artificial Intelligence, remains a subject of debate. This work explores the historical precedents of technological breakthroughs, examining whether Artificial Intelligence can achieve a comparable impact, and it delves into various ethical frameworks that shape the perception and development of Artificial Intelligence. Additionally, it considers the societal, technical, and regulatory challenges that must be addressed for Artificial Intelligence to become a catalyst for global change. We also examine not only the strategies and methodologies that could lead to transformative Artificial Intelligence but also the barriers that could ultimately make these goals unattainable. We end with a critical inquiry into whether reaching a transformative Artificial Intelligence might compel humanity to adopt an entirely new ethical approach, tailored to the complexities of advanced Artificial Intelligence. By addressing the ethical, social, and scientific dimensions of Artificial Intelligence's development, this work contributes to the broader discourse on the long-term implications of Artificial Intelligence and its capacity to drive civilization toward a new era of progress or, conversely, exacerbate existing inequalities and risks.
comment: 100 pages, 6 Figures, 2 Tables
☆ Position-aware Guided Point Cloud Completion with CLIP Model AAAI25
Point cloud completion aims to recover partial geometric and topological shapes caused by equipment defects or limited viewpoints. Current methods either solely rely on the 3D coordinates of the point cloud to complete it or incorporate additional images with well-calibrated intrinsic parameters to guide the geometric estimation of the missing parts. Although these methods have achieved excellent performance by directly predicting the location of complete points, the extracted features lack fine-grained information regarding the location of the missing area. To address this issue, we propose a rapid and efficient method to expand an unimodal framework into a multimodal framework. This approach incorporates a position-aware module designed to enhance the spatial information of the missing parts through a weighted map learning mechanism. In addition, we establish a Point-Text-Image triplet corpus PCI-TI and MVP-TI based on the existing unimodal point cloud completion dataset and use the pre-trained vision-language model CLIP to provide richer detail information for 3D shapes, thereby enhancing performance. Extensive quantitative and qualitative experiments demonstrate that our method outperforms state-of-the-art point cloud completion methods.
comment: Accepted by AAAI25
☆ FLIP: Flow-Centric Generative Planning for General-Purpose Manipulation Tasks
We aim to develop a model-based planning framework for world models that can be scaled with increasing model and data budgets for general-purpose manipulation tasks with only language and vision inputs. To this end, we present FLow-centric generative Planning (FLIP), a model-based planning algorithm on visual space that features three key modules: 1. a multi-modal flow generation model as the general-purpose action proposal module; 2. a flow-conditioned video generation model as the dynamics module; and 3. a vision-language representation learning model as the value module. Given an initial image and language instruction as the goal, FLIP can progressively search for long-horizon flow and video plans that maximize the discounted return to accomplish the task. FLIP is able to synthesize long-horizon plans across objects, robots, and tasks with image flows as the general action representation, and the dense flow information also provides rich guidance for long-horizon video generation. In addition, the synthesized flow and video plans can guide the training of low-level control policies for robot execution. Experiments on diverse benchmarks demonstrate that FLIP can improve both the success rates and quality of long-horizon video plan synthesis and has the interactive world model property, opening up wider applications for future works.
☆ Large Language Models for Scholarly Ontology Generation: An Extensive Analysis in the Engineering Field
Ontologies of research topics are crucial for structuring scientific knowledge, enabling scientists to navigate vast amounts of research, and forming the backbone of intelligent systems such as search engines and recommendation systems. However, manual creation of these ontologies is expensive, slow, and often results in outdated and overly general representations. As a solution, researchers have been investigating ways to automate or semi-automate the process of generating these ontologies. This paper offers a comprehensive analysis of the ability of large language models (LLMs) to identify semantic relationships between different research topics, which is a critical step in the development of such ontologies. To this end, we developed a gold standard based on the IEEE Thesaurus to evaluate the task of identifying four types of relationships between pairs of topics: broader, narrower, same-as, and other. Our study evaluates the performance of seventeen LLMs, which differ in scale, accessibility (open vs. proprietary), and model type (full vs. quantised), while also assessing four zero-shot reasoning strategies. Several models have achieved outstanding results, including Mixtral-8x7B, Dolphin-Mistral-7B, and Claude 3 Sonnet, with F1-scores of 0.847, 0.920, and 0.967, respectively. Furthermore, our findings demonstrate that smaller, quantised models, when optimised through prompt engineering, can deliver performance comparable to much larger proprietary models, while requiring significantly fewer computational resources.
comment: submitted to Information Processing & Management
☆ Dynamic Modality-Camera Invariant Clustering for Unsupervised Visible-Infrared Person Re-identification
Unsupervised learning visible-infrared person re-identification (USL-VI-ReID) offers a more flexible and cost-effective alternative compared to supervised methods. This field has gained increasing attention due to its promising potential. Existing methods simply cluster modality-specific samples and employ strong association techniques to achieve instance-to-cluster or cluster-to-cluster cross-modality associations. However, they ignore cross-camera differences, leading to noticeable issues with excessive splitting of identities. Consequently, this undermines the accuracy and reliability of cross-modal associations. To address these issues, we propose a novel Dynamic Modality-Camera Invariant Clustering (DMIC) framework for USL-VI-ReID. Specifically, our DMIC naturally integrates Modality-Camera Invariant Expansion (MIE), Dynamic Neighborhood Clustering (DNC) and Hybrid Modality Contrastive Learning (HMCL) into a unified framework, which eliminates both the cross-modality and cross-camera discrepancies in clustering. MIE fuses inter-modal and inter-camera distance coding to bridge the gaps between modalities and cameras at the clustering level. DNC employs two dynamic search strategies to refine the network's optimization objective, transitioning from improving discriminability to enhancing cross-modal and cross-camera generalizability. Moreover, HMCL is designed to optimize instance-level and cluster-level distributions. Memories for intra-modality and inter-modality training are updated using randomly selected samples, facilitating real-time exploration of modality-invariant representations. Extensive experiments have demonstrated that our DMIC addresses the limitations present in current clustering approaches and achieve competitive performance, which significantly reduces the performance gap with supervised methods.
☆ Hierarchical Classification for Automated Image Annotation of Coral Reef Benthic Structures NeurIPS 2024
Automated benthic image annotation is crucial to efficiently monitor and protect coral reefs against climate change. Current machine learning approaches fail to capture the hierarchical nature of benthic organisms covering reef substrata, i.e., coral taxonomic levels and health condition. To address this limitation, we propose to annotate benthic images using hierarchical classification. Experiments on a custom dataset from a Northeast Brazilian coral reef show that our approach outperforms flat classifiers, improving both F1 and hierarchical F1 scores by approximately 2\% across varying amounts of training data. In addition, this hierarchical method aligns more closely with ecological objectives.
comment: Poster at Tackling Climate Change with Machine Learning: workshop at NeurIPS 2024
☆ Generate Any Scene: Evaluating and Improving Text-to-Vision Generation with Scene Graph Programming
DALL-E and Sora have gained attention by producing implausible images, such as "astronauts riding a horse in space." Despite the proliferation of text-to-vision models that have inundated the internet with synthetic visuals, from images to 3D assets, current benchmarks predominantly evaluate these models on real-world scenes paired with captions. We introduce Generate Any Scene, a framework that systematically enumerates scene graphs representing a vast array of visual scenes, spanning realistic to imaginative compositions. Generate Any Scene leverages 'scene graph programming', a method for dynamically constructing scene graphs of varying complexity from a structured taxonomy of visual elements. This taxonomy includes numerous objects, attributes, and relations, enabling the synthesis of an almost infinite variety of scene graphs. Using these structured representations, Generate Any Scene translates each scene graph into a caption, enabling scalable evaluation of text-to-vision models through standard metrics. We conduct extensive evaluations across multiple text-to-image, text-to-video, and text-to-3D models, presenting key findings on model performance. We find that DiT-backbone text-to-image models align more closely with input captions than UNet-backbone models. Text-to-video models struggle with balancing dynamics and consistency, while both text-to-video and text-to-3D models show notable gaps in human preference alignment. We demonstrate the effectiveness of Generate Any Scene by conducting three practical applications leveraging captions generated by Generate Any Scene: 1) a self-improving framework where models iteratively enhance their performance using generated data, 2) a distillation process to transfer specific strengths from proprietary models to open-source counterparts, and 3) improvements in content moderation by identifying and generating challenging synthetic data.
☆ SAFIRE: Segment Any Forged Image Region AAAI 2025
Most techniques approach the problem of image forgery localization as a binary segmentation task, training neural networks to label original areas as 0 and forged areas as 1. In contrast, we tackle this issue from a more fundamental perspective by partitioning images according to their originating sources. To this end, we propose Segment Any Forged Image Region (SAFIRE), which solves forgery localization using point prompting. Each point on an image is used to segment the source region containing itself. This allows us to partition images into multiple source regions, a capability achieved for the first time. Additionally, rather than memorizing certain forgery traces, SAFIRE naturally focuses on uniform characteristics within each source region. This approach leads to more stable and effective learning, achieving superior performance in both the new task and the traditional binary forgery localization.
comment: Accepted at AAAI 2025. Code is available at: https://github.com/mjkwon2021/SAFIRE
☆ Semantic Scene Completion Based 3D Traversability Estimation for Off-Road Terrains
Off-road environments present significant challenges for autonomous ground vehicles due to the absence of structured roads and the presence of complex obstacles, such as uneven terrain, vegetation, and occlusions. Traditional perception algorithms, designed primarily for structured environments, often fail under these conditions, leading to inaccurate traversability estimations. In this paper, ORDformer, a novel multimodal method that combines LiDAR point clouds with monocular images, is proposed to generate dense traversable occupancy predictions from a forward-facing perspective. By integrating multimodal data, environmental feature extraction is enhanced, which is crucial for accurate occupancy estimation in complex terrains. Furthermore, RELLIS-OCC, a dataset with 3D traversable occupancy annotations, is introduced, incorporating geometric features such as step height, slope, and unevenness. Through a comprehensive analysis of vehicle obstacle-crossing conditions and the incorporation of vehicle body structure constraints, four traversability cost labels are generated: lethal, medium-cost, low-cost, and free. Experimental results demonstrate that ORDformer outperforms existing approaches in 3D traversable area recognition, particularly in off-road environments with irregular geometries and partial occlusions. Specifically, ORDformer achieves over a 20\% improvement in scene completion IoU compared to other models. The proposed framework is scalable and adaptable to various vehicle platforms, allowing for adjustments to occupancy grid parameters and the integration of advanced dynamic models for traversability cost estimation.
comment: 12 pages,14 figures
☆ From communities to interpretable network and word embedding: an unified approach
Modelling information from complex systems such as humans social interaction or words co-occurrences in our languages can help to understand how these systems are organized and function. Such systems can be modelled by networks, and network theory provides a useful set of methods to analyze them. Among these methods, graph embedding is a powerful tool to summarize the interactions and topology of a network in a vectorized feature space. When used in input of machine learning algorithms, embedding vectors help with common graph problems such as link prediction, graph matching, etc. Word embedding has the goal of representing the sense of words, extracting it from large text corpora. Despite differences in the structure of information in input of embedding algorithms, many graph embedding approaches are adapted and inspired from methods in NLP. Limits of these methods are observed in both domains. Most of these methods require long and resource greedy training. Another downside to most methods is that they are black-box, from which understanding how the information is structured is rather complex. Interpretability of a model allows understanding how the vector space is structured without the need for external information, and thus can be audited more easily. With both these limitations in mind, we propose a novel framework to efficiently embed network vertices in an interpretable vector space. Our Lower Dimension Bipartite Framework (LDBGF) leverages the bipartite projection of a network using cliques to reduce dimensionality. Along with LDBGF, we introduce two implementations of this framework that rely on communities instead of cliques: SINr-NR and SINr-MF. We show that SINr-MF can perform well on classical graphs and SINr-NR can produce high-quality graph and word embeddings that are interpretable and stable across runs.
☆ Towards Automated Algebraic Multigrid Preconditioner Design Using Genetic Programming for Large-Scale Laser Beam Welding Simulations
Multigrid methods are asymptotically optimal algorithms ideal for large-scale simulations. But, they require making numerous algorithmic choices that significantly influence their efficiency. Unlike recent approaches that learn optimal multigrid components using machine learning techniques, we adopt a complementary strategy here, employing evolutionary algorithms to construct efficient multigrid cycles from available individual components. This technology is applied to finite element simulations of the laser beam welding process. The thermo-elastic behavior is described by a coupled system of time-dependent thermo-elasticity equations, leading to nonlinear and ill-conditioned systems. The nonlinearity is addressed using Newton's method, and iterative solvers are accelerated with an algebraic multigrid (AMG) preconditioner using hypre BoomerAMG interfaced via PETSc. This is applied as a monolithic solver for the coupled equations. To further enhance solver efficiency, flexible AMG cycles are introduced, extending traditional cycle types with level-specific smoothing sequences and non-recursive cycling patterns. These are automatically generated using genetic programming, guided by a context-free grammar containing AMG rules. Numerical experiments demonstrate the potential of these approaches to improve solver performance in large-scale laser beam welding simulations.
☆ Auto-Generating Earnings Report Analysis via a Financial-Augmented LLM
Financial analysis heavily relies on the evaluation of earnings reports to gain insights into company performance. Traditional generation of these reports requires extensive financial expertise and is time-consuming. With the impressive progress in Large Language Models (LLMs), a wide variety of financially focused LLMs has emerged, addressing tasks like sentiment analysis and entity recognition in the financial domain. This paper presents a novel challenge: developing an LLM specifically for automating the generation of earnings reports analysis. Our methodology involves an in-depth analysis of existing earnings reports followed by a unique approach to fine-tune an LLM for this purpose. This approach combines retrieval augmentation and the generation of instruction-based data, specifically tailored for the financial sector, to enhance the LLM's performance. With extensive financial documents, we construct financial instruction data, enabling the refined adaptation of our LLM to financial contexts. Preliminary results indicate that our augmented LLM outperforms general open-source models and rivals commercial counterparts like GPT-3.5 in financial applications. Our research paves the way for streamlined and insightful automation in financial report generation, marking a significant stride in the field of financial analysis.
comment: 8 pages, 1 figure, 4 tables
☆ Can Graph Neural Networks Learn Language with Extremely Weak Text Supervision?
While great success has been achieved in building vision models with Contrastive Language-Image Pre-training (CLIP) over Internet-scale image-text pairs, building transferable Graph Neural Networks (GNNs) with CLIP pipeline is challenging because of three fundamental issues: the scarcity of labeled data and text supervision, different levels of downstream tasks, and the conceptual gaps between domains. In this work, to address these issues, we leverage multi-modal prompt learning to effectively adapt pre-trained GNN to downstream tasks and data, given only a few semantically labeled samples, each with extremely weak text supervision. Our new paradigm embeds the graphs directly in the same space as the Large Language Models (LLMs) by learning both graph prompts and text prompts simultaneously. To accomplish this, we improve state-of-the-art graph prompt method, and then propose the first graph-language multi-modal prompt learning approach for exploiting the knowledge in pre-trained models. Notably, due to the insufficient supervision for fine-tuning, in our paradigm, the pre-trained GNN and the LLM are kept frozen, so the learnable parameters are much fewer than fine-tuning any pre-trained model. Through extensive experiments on real-world datasets, we demonstrate the superior performance of our paradigm in few-shot, multi-task-level, and cross-domain settings. Moreover, we build the first CLIP-style zero-shot classification prototype that can generalize GNNs to unseen classes with extremely weak text supervision.
comment: Preprint, 26 pages
☆ Antelope: Potent and Concealed Jailbreak Attack Strategy
Due to the remarkable generative potential of diffusion-based models, numerous researches have investigated jailbreak attacks targeting these frameworks. A particularly concerning threat within image models is the generation of Not-Safe-for-Work (NSFW) content. Despite the implementation of security filters, numerous efforts continue to explore ways to circumvent these safeguards. Current attack methodologies primarily encompass adversarial prompt engineering or concept obfuscation, yet they frequently suffer from slow search efficiency, conspicuous attack characteristics and poor alignment with targets. To overcome these challenges, we propose Antelope, a more robust and covert jailbreak attack strategy designed to expose security vulnerabilities inherent in generative models. Specifically, Antelope leverages the confusion of sensitive concepts with similar ones, facilitates searches in the semantically adjacent space of these related concepts and aligns them with the target imagery, thereby generating sensitive images that are consistent with the target and capable of evading detection. Besides, we successfully exploit the transferability of model-based attacks to penetrate online black-box services. Experimental evaluations demonstrate that Antelope outperforms existing baselines across multiple defensive mechanisms, underscoring its efficacy and versatility.
☆ A Review of Intelligent Device Fault Diagnosis Technologies Based on Machine Vision
This paper provides a comprehensive review of mechanical equipment fault diagnosis methods, focusing on the advancements brought by Transformer-based models. It details the structure, working principles, and benefits of Transformers, particularly their self-attention mechanism and parallel computation capabilities, which have propelled their widespread application in natural language processing and computer vision. The discussion highlights key Transformer model variants, such as Vision Transformers (ViT) and their extensions, which leverage self-attention to improve accuracy and efficiency in visual tasks. Furthermore, the paper examines the application of Transformer-based approaches in intelligent fault diagnosis for mechanical systems, showcasing their superior ability to extract and recognize patterns from complex sensor data for precise fault identification. Despite these advancements, challenges remain, including the reliance on extensive labeled datasets, significant computational demands, and difficulties in deploying models on resource-limited devices. To address these limitations, the paper proposes future research directions, such as developing lightweight Transformer architectures, integrating multimodal data sources, and enhancing adaptability to diverse operational conditions. These efforts aim to further expand the application of Transformer-based methods in mechanical fault diagnosis, making them more robust, efficient, and suitable for real-world industrial environments.
comment: 9 pages, This paper has been accepted for publication at RICAI 2024
☆ How to Weight Multitask Finetuning? Fast Previews via Bayesian Model-Merging
When finetuning multiple tasks altogether, it is important to carefully weigh them to get a good performance, but searching for good weights can be difficult and costly. Here, we propose to aid the search with fast previews to quickly get a rough idea of different reweighting options. We use model merging to create previews by simply reusing and averaging parameters of models trained on each task separately (no retraining required). To improve the quality of previews, we propose a Bayesian approach to design new merging strategies by using more flexible posteriors. We validate our findings on vision and natural-language transformers. Our work shows the benefits of model merging via Bayes to improve multitask finetuning.
☆ A Survey on Private Transformer Inference
Transformer models have revolutionized AI, enabling applications like content generation and sentiment analysis. However, their use in Machine Learning as a Service (MLaaS) raises significant privacy concerns, as centralized servers process sensitive user data. Private Transformer Inference (PTI) addresses these issues using cryptographic techniques such as Secure Multi-Party Computation (MPC) and Homomorphic Encryption (HE), enabling secure model inference without exposing inputs or models. This paper reviews recent advancements in PTI, analyzing state-of-the-art solutions, their challenges, and potential improvements. We also propose evaluation guidelines to assess resource efficiency and privacy guarantees, aiming to bridge the gap between high-performance inference and data privacy.
comment: The manuscript is still being revised and will be continuously updated in the future
☆ AGMixup: Adaptive Graph Mixup for Semi-supervised Node Classification AAAI 2025
Mixup is a data augmentation technique that enhances model generalization by interpolating between data points using a mixing ratio $\lambda$ in the image domain. Recently, the concept of mixup has been adapted to the graph domain through node-centric interpolations. However, these approaches often fail to address the complexity of interconnected relationships, potentially damaging the graph's natural topology and undermining node interactions. Furthermore, current graph mixup methods employ a one-size-fits-all strategy with a randomly sampled $\lambda$ for all mixup pairs, ignoring the diverse needs of different pairs. This paper proposes an Adaptive Graph Mixup (AGMixup) framework for semi-supervised node classification. AGMixup introduces a subgraph-centric approach, which treats each subgraph similarly to how images are handled in Euclidean domains, thus facilitating a more natural integration of mixup into graph-based learning. We also propose an adaptive mechanism to tune the mixing ratio $\lambda$ for diverse mixup pairs, guided by the contextual similarity and uncertainty of the involved subgraphs. Extensive experiments across seven datasets on semi-supervised node classification benchmarks demonstrate AGMixup's superiority over state-of-the-art graph mixup methods. Source codes are available at \url{https://github.com/WeigangLu/AGMixup}.
comment: Accepted by AAAI 2025
☆ Wasserstein Distance Rivals Kullback-Leibler Divergence for Knowledge Distillation NeurIPS 2024
Since pioneering work of Hinton et al., knowledge distillation based on Kullback-Leibler Divergence (KL-Div) has been predominant, and recently its variants have achieved compelling performance. However, KL-Div only compares probabilities of the corresponding category between the teacher and student while lacking a mechanism for cross-category comparison. Besides, KL-Div is problematic when applied to intermediate layers, as it cannot handle non-overlapping distributions and is unaware of geometry of the underlying manifold. To address these downsides, we propose a methodology of Wasserstein Distance (WD) based knowledge distillation. Specifically, we propose a logit distillation method called WKD-L based on discrete WD, which performs cross-category comparison of probabilities and thus can explicitly leverage rich interrelations among categories. Moreover, we introduce a feature distillation method called WKD-F, which uses a parametric method for modeling feature distributions and adopts continuous WD for transferring knowledge from intermediate layers. Comprehensive evaluations on image classification and object detection have shown (1) for logit distillation WKD-L outperforms very strong KL-Div variants; (2) for feature distillation WKD-F is superior to the KL-Div counterparts and state-of-the-art competitors. The source code is available at https://peihuali.org/WKD
comment: Accepted to NeurIPS 2024. Equal contribution from first two authors
☆ Learn How to Query from Unlabeled Data Streams in Federated Learning
Federated learning (FL) enables collaborative learning among decentralized clients while safeguarding the privacy of their local data. Existing studies on FL typically assume offline labeled data available at each client when the training starts. Nevertheless, the training data in practice often arrive at clients in a streaming fashion without ground-truth labels. Given the expensive annotation cost, it is critical to identify a subset of informative samples for labeling on clients. However, selecting samples locally while accommodating the global training objective presents a challenge unique to FL. In this work, we tackle this conundrum by framing the data querying process in FL as a collaborative decentralized decision-making problem and proposing an effective solution named LeaDQ, which leverages multi-agent reinforcement learning algorithms. In particular, under the implicit guidance from global information, LeaDQ effectively learns the local policies for distributed clients and steers them towards selecting samples that can enhance the global model's accuracy. Extensive simulations on image and text tasks show that LeaDQ advances the model performance in various FL scenarios, outperforming the benchmarking algorithms.
☆ Intelligent Electric Power Steering: Artificial Intelligence Integration Enhances Vehicle Safety and Performance
Electric Power Steering (EPS) systems utilize electric motors to aid users in steering their vehicles, which provide additional precise control and reduced energy consumption compared to traditional hydraulic systems. EPS technology provides safety,control and efficiency.. This paper explains the integration of Artificial Intelligence (AI) into Electric Power Steering (EPS) systems, focusing on its role in enhancing the safety, and adaptability across diverse driving conditions. We explore significant development in AI-driven EPS, including predictive control algorithms, adaptive torque management systems, and data-driven diagnostics. The paper presents case studies of AI applications in EPS, such as Lane centering control (LCC), Automated Parking Systems, and Autonomous Vehicle Steering, while considering the challenges, limitations, and future prospects of this technology. This article discusses current developments in AI-driven EPS, emphasizing on the benefits of improved safety, adaptive control, and predictive maintenance. Challenges in integrating AI in EPS systems. This paper addresses cybersecurity risks, ethical concerns, and technical limitations,, along with next steps for research and implementation in autonomous, and connected vehicles.
comment: IEEE Summit on Reliability, Availability and Serviceability, 2024
☆ DiffRaman: A Conditional Latent Denoising Diffusion Probabilistic Model for Bacterial Raman Spectroscopy Identification Under Limited Data Conditions
Raman spectroscopy has attracted significant attention in various biochemical detection fields, especially in the rapid identification of pathogenic bacteria. The integration of this technology with deep learning to facilitate automated bacterial Raman spectroscopy diagnosis has emerged as a key focus in recent research. However, the diagnostic performance of existing deep learning methods largely depends on a sufficient dataset, and in scenarios where there is a limited availability of Raman spectroscopy data, it is inadequate to fully optimize the numerous parameters of deep neural networks. To address these challenges, this paper proposes a data generation method utilizing deep generative models to expand the data volume and enhance the recognition accuracy of bacterial Raman spectra. Specifically, we introduce DiffRaman, a conditional latent denoising diffusion probability model for Raman spectra generation. Experimental results demonstrate that synthetic bacterial Raman spectra generated by DiffRaman can effectively emulate real experimental spectra, thereby enhancing the performance of diagnostic models, especially under conditions of limited data. Furthermore, compared to existing generative models, the proposed DiffRaman offers improvements in both generation quality and computational efficiency. Our DiffRaman approach offers a well-suited solution for automated bacteria Raman spectroscopy diagnosis in data-scarce scenarios, offering new insights into alleviating the labor of spectroscopic measurements and enhancing rare bacteria identification.
☆ Evil twins are not that evil: Qualitative insights into machine-generated prompts
It has been widely observed that language models (LMs) respond in predictable ways to algorithmically generated prompts that are seemingly unintelligible. This is both a sign that we lack a full understanding of how LMs work, and a practical challenge, because opaqueness can be exploited for harmful uses of LMs, such as jailbreaking. We present the first thorough analysis of opaque machine-generated prompts, or autoprompts, pertaining to 3 LMs of different sizes and families. We find that machine-generated prompts are characterized by a last token that is often intelligible and strongly affects the generation. A small but consistent proportion of the previous tokens are fillers that probably appear in the prompt as a by-product of the fact that the optimization process fixes the number of tokens. The remaining tokens tend to have at least a loose semantic relation with the generation, although they do not engage in well-formed syntactic relations with it. We find moreover that some of the ablations we applied to machine-generated prompts can also be applied to natural language sequences, leading to similar behavior, suggesting that autoprompts are a direct consequence of the way in which LMs process linguistic inputs in general.
☆ Dense Depth from Event Focal Stack WACV2025
We propose a method for dense depth estimation from an event stream generated when sweeping the focal plane of the driving lens attached to an event camera. In this method, a depth map is inferred from an ``event focal stack'' composed of the event stream using a convolutional neural network trained with synthesized event focal stacks. The synthesized event stream is created from a focal stack generated by Blender for any arbitrary 3D scene. This allows for training on scenes with diverse structures. Additionally, we explored methods to eliminate the domain gap between real event streams and synthetic event streams. Our method demonstrates superior performance over a depth-from-defocus method in the image domain on synthetic and real datasets.
comment: Accepted at WACV2025
☆ LatentSpeech: Latent Diffusion for Text-To-Speech Generation
Diffusion-based Generative AI gains significant attention for its superior performance over other generative techniques like Generative Adversarial Networks and Variational Autoencoders. While it has achieved notable advancements in fields such as computer vision and natural language processing, their application in speech generation remains under-explored. Mainstream Text-to-Speech systems primarily map outputs to Mel-Spectrograms in the spectral space, leading to high computational loads due to the sparsity of MelSpecs. To address these limitations, we propose LatentSpeech, a novel TTS generation approach utilizing latent diffusion models. By using latent embeddings as the intermediate representation, LatentSpeech reduces the target dimension to 5% of what is required for MelSpecs, simplifying the processing for the TTS encoder and vocoder and enabling efficient high-quality speech generation. This study marks the first integration of latent diffusion models in TTS, enhancing the accuracy and naturalness of generated speech. Experimental results on benchmark datasets demonstrate that LatentSpeech achieves a 25% improvement in Word Error Rate and a 24% improvement in Mel Cepstral Distortion compared to existing models, with further improvements rising to 49.5% and 26%, respectively, with additional training data. These findings highlight the potential of LatentSpeech to advance the state-of-the-art in TTS technology
☆ Aligner-Guided Training Paradigm: Advancing Text-to-Speech Models with Aligner Guided Duration
Recent advancements in text-to-speech (TTS) systems, such as FastSpeech and StyleSpeech, have significantly improved speech generation quality. However, these models often rely on duration generated by external tools like the Montreal Forced Aligner, which can be time-consuming and lack flexibility. The importance of accurate duration is often underestimated, despite their crucial role in achieving natural prosody and intelligibility. To address these limitations, we propose a novel Aligner-Guided Training Paradigm that prioritizes accurate duration labelling by training an aligner before the TTS model. This approach reduces dependence on external tools and enhances alignment accuracy. We further explore the impact of different acoustic features, including Mel-Spectrograms, MFCCs, and latent features, on TTS model performance. Our experimental results show that aligner-guided duration labelling can achieve up to a 16\% improvement in word error rate and significantly enhance phoneme and tone alignment. These findings highlight the effectiveness of our approach in optimizing TTS systems for more natural and intelligible speech generation.
☆ Unseen Horizons: Unveiling the Real Capability of LLM Code Generation Beyond the Familiar
Recently, large language models (LLMs) have shown strong potential in code generation tasks. However, there are still gaps before they can be fully applied in actual software development processes. Accurately assessing the code generation capabilities of large language models has become an important basis for evaluating and improving the models. Some existing works have constructed datasets to evaluate the capabilities of these models. However, the current evaluation process may encounter the illusion of "Specialist in Familiarity", primarily due to three gaps: the exposure of target code, case timeliness, and dependency availability. The fundamental reason for these gaps is that the code in current datasets may have been extensively exposed and exercised during the training phase, and due to the continuous training and development of LLM, their timeliness has been severely compromised. The key to solve the problem is to, as much as possible, evaluate the LLMs using code that they have not encountered before. Thus, the fundamental idea in this paper is to draw on the concept of code obfuscation, changing code at different levels while ensuring the functionality and output. To this end, we build a code-obfuscation based benchmark OBFUSEVAL. We first collect 1,354 raw cases from five real-world projects, including function description and code. Then we use three-level strategy (symbol, structure and semantic) to obfuscate descriptions, code and context dependencies. We evaluate four LLMs on OBFU- SEVAL and compared the effectiveness of different obfuscation strategy. We use official test suites of these projects to evaluate the generated code. The results show that after obfuscation, the average decrease ratio of test pass rate can up to 62.5%.
comment: Large Language Model,Code Generation Capability,Code Dataset
☆ Adversarial Vulnerabilities in Large Language Models for Time Series Forecasting
Large Language Models (LLMs) have recently demonstrated significant potential in the field of time series forecasting, offering impressive capabilities in handling complex temporal data. However, their robustness and reliability in real-world applications remain under-explored, particularly concerning their susceptibility to adversarial attacks. In this paper, we introduce a targeted adversarial attack framework for LLM-based time series forecasting. By employing both gradient-free and black-box optimization methods, we generate minimal yet highly effective perturbations that significantly degrade the forecasting accuracy across multiple datasets and LLM architectures. Our experiments, which include models like TimeGPT and LLM-Time with GPT-3.5, GPT-4, LLaMa, and Mistral, show that adversarial attacks lead to much more severe performance degradation than random noise, and demonstrate the broad effectiveness of our attacks across different LLMs. The results underscore the critical vulnerabilities of LLMs in time series forecasting, highlighting the need for robust defense mechanisms to ensure their reliable deployment in practical applications.
comment: 11 pages, 5 figures
☆ What You See Is Not Always What You Get: An Empirical Study of Code Comprehension by Large Language Models
Recent studies have demonstrated outstanding capabilities of large language models (LLMs) in software engineering domain, covering numerous tasks such as code generation and comprehension. While the benefit of LLMs for coding task is well noted, it is perceived that LLMs are vulnerable to adversarial attacks. In this paper, we study the specific LLM vulnerability to imperceptible character attacks, a type of prompt-injection attack that uses special characters to befuddle an LLM whilst keeping the attack hidden to human eyes. We devise four categories of attacks and investigate their effects on the performance outcomes of tasks relating to code analysis and code comprehension. Two generations of ChatGPT are included to evaluate the impact of advancements made to contemporary models. Our experimental design consisted of comparing perturbed and unperturbed code snippets and evaluating two performance outcomes, which are model confidence using log probabilities of response, and correctness of response. We conclude that earlier version of ChatGPT exhibits a strong negative linear correlation between the amount of perturbation and the performance outcomes, while the recent ChatGPT presents a strong negative correlation between the presence of perturbation and performance outcomes, but no valid correlational relationship between perturbation budget and performance outcomes. We anticipate this work contributes to an in-depth understanding of leveraging LLMs for coding tasks. It is suggested future research should delve into how to create LLMs that can return a correct response even if the prompt exhibits perturbations.
☆ Multilingual LLMs Inherently Reward In-Language Time-Sensitive Semantic Alignment for Low-Resource Languages
The unwavering disparity in labeled resources between resource-rich languages and those considered low-resource remains a significant impediment for Large Language Models (LLMs). Recent strides in cross-lingual in-context learning (X-ICL), mainly through semantically aligned examples retrieved from multilingual pre-trained transformers, have shown promise in mitigating this issue. However, our investigation reveals that LLMs intrinsically reward in-language semantically aligned cross-lingual instances over direct cross-lingual semantic alignments, with a pronounced disparity in handling time-sensitive queries in the X-ICL setup. Such queries demand sound temporal reasoning ability from LLMs, yet the advancements have predominantly focused on English. This study aims to bridge this gap by improving temporal reasoning capabilities in low-resource languages. To this end, we introduce mTEMPREASON a temporal reasoning dataset aimed at the varied degrees of low-resource languages and propose Cross-Lingual Time-Sensitive Semantic Alignment (CLiTSSA), a novel method to improve temporal reasoning in these contexts. To facilitate this, we construct an extension of mTEMPREASON comprising pairs of parallel cross-language temporal queries along with their anticipated in-language semantic similarity scores. Our empirical evidence underscores the superior performance of CLiTSSA compared to established baselines across three languages - Romanian, German, and French, encompassing three temporal tasks and including a diverse set of four contemporaneous LLMs. This marks a significant step forward in addressing resource disparity in the context of temporal reasoning across languages.
☆ Non-Myopic Multi-Objective Bayesian Optimization
We consider the problem of finite-horizon sequential experimental design to solve multi-objective optimization (MOO) of expensive black-box objective functions. This problem arises in many real-world applications, including materials design, where we have a small resource budget to make and evaluate candidate materials in the lab. We solve this problem using the framework of Bayesian optimization (BO) and propose the first set of non-myopic methods for MOO problems. Prior work on non-myopic BO for single-objective problems relies on the Bellman optimality principle to handle the lookahead reasoning process. However, this principle does not hold for most MOO problems because the reward function needs to satisfy some conditions: scalar variable, monotonicity, and additivity. We address this challenge by using hypervolume improvement (HVI) as our scalarization approach, which allows us to use a lower-bound on the Bellman equation to approximate the finite-horizon using a batch expected hypervolume improvement (EHVI) acquisition function (AF) for MOO. Our formulation naturally allows us to use other improvement-based scalarizations and compare their efficacy to HVI. We derive three non-myopic AFs for MOBO: 1) the Nested AF, which is based on the exact computation of the lower bound, 2) the Joint AF, which is a lower bound on the nested AF, and 3) the BINOM AF, which is a fast and approximate variant based on batch multi-objective acquisition functions. Our experiments on multiple diverse real-world MO problems demonstrate that our non-myopic AFs substantially improve performance over the existing myopic AFs for MOBO.
☆ How to select slices for annotation to train best-performing deep learning segmentation models for cross-sectional medical images?
Automated segmentation of medical images highly depends on the availability of accurate manual image annotations. Such annotations are very time-consuming and costly to generate, and often require specialized expertise, particularly for cross-sectional images which contain many slices for each patient. It is crucial to ensure the best use of annotation resources. In this paper, we systematically answer the question of how to select slices of cross-sectional medical images in order to maximize performance of the resulting deep learning segmentation models. We conducted experiments on 4 medical imaging segmentation tasks with varying annotation budgets, numbers of annotated cases, numbers of annotated slices per volume, slice selection techniques, and mask interpolations. We found that: 1) It is almost always preferable to annotate fewer slices per volume and more volumes given an annotation budget. 2) Selecting slices for annotation by unsupervised active learning (UAL) is not superior to selecting slices randomly or at fixed intervals, provided that each volume is allocated the same number of annotated slices. 3) Interpolating masks between annotated slices rarely enhances model performance, with exceptions of some specific configuration for 3D models.
☆ Using Large Language Models for Parametric Shape Optimization
Recent advanced large language models (LLMs) have showcased their emergent capability of in-context learning, facilitating intelligent decision-making through natural language prompts without retraining. This new machine learning paradigm has shown promise in various fields, including general control and optimization problems. Inspired by these advancements, we explore the potential of LLMs for a specific and essential engineering task: parametric shape optimization (PSO). We develop an optimization framework, LLM-PSO, that leverages an LLM to determine the optimal shape of parameterized engineering designs in the spirit of evolutionary strategies. Utilizing the ``Claude 3.5 Sonnet'' LLM, we evaluate LLM-PSO on two benchmark flow optimization problems, specifically aiming to identify drag-minimizing profiles for 1) a two-dimensional airfoil in laminar flow, and 2) a three-dimensional axisymmetric body in Stokes flow. In both cases, LLM-PSO successfully identifies optimal shapes in agreement with benchmark solutions. Besides, it generally converges faster than other classical optimization algorithms. Our preliminary exploration may inspire further investigations into harnessing LLMs for shape optimization and engineering design more broadly.
☆ DialogAgent: An Auto-engagement Agent for Code Question Answering Data Production
Large Language Models (LLMs) have become increasingly integral to enhancing developer productivity, particularly in code generation, comprehension, and repair tasks. However, fine-tuning these models with high-quality, real-world data is challenging due to privacy concerns and the lack of accessible, labeled datasets. In this paper, we present DialogAgent, an automated tool for generating synthetic training data that closely mimics real developer interactions within Integrated Development Environments (IDEs). DialogAgent enables the production of diverse, high-fidelity query-response pairs by simulating multi-turn dialogues and contextual behaviors observed in real-world programming scenarios. The tool significantly reduces the reliance on manual data generation, increasing efficiency by 4.8 times compared to traditional methods. Our experiments and online deployment demonstrate substantial improvements in model performance for code-related question-answering tasks: the acceptance rate of responses generated by our in-house model is improved by 33%, after training on synthesized data generated by DialogAgent.
☆ Repository-Level Graph Representation Learning for Enhanced Security Patch Detection ICSE 2025
Software vendors often silently release security patches without providing sufficient advisories (e.g., Common Vulnerabilities and Exposures) or delayed updates via resources (e.g., National Vulnerability Database). Therefore, it has become crucial to detect these security patches to ensure secure software maintenance. However, existing methods face the following challenges: (1) They primarily focus on the information within the patches themselves, overlooking the complex dependencies in the repository. (2) Security patches typically involve multiple functions and files, increasing the difficulty in well learning the representations. To alleviate the above challenges, this paper proposes a Repository-level Security Patch Detection framework named RepoSPD, which comprises three key components: 1) a repository-level graph construction, RepoCPG, which represents software patches by merging pre-patch and post-patch source code at the repository level; 2) a structure-aware patch representation, which fuses the graph and sequence branch and aims at comprehending the relationship among multiple code changes; 3) progressive learning, which facilitates the model in balancing semantic and structural information. To evaluate RepoSPD, we employ two widely-used datasets in security patch detection: SPI-DB and PatchDB. We further extend these datasets to the repository level, incorporating a total of 20,238 and 28,781 versions of repository in C/C++ programming languages, respectively, denoted as SPI-DB* and PatchDB*. We compare RepoSPD with six existing security patch detection methods and five static tools. Our experimental results demonstrate that RepoSPD outperforms the state-of-the-art baseline, with improvements of 11.90%, and 3.10% in terms of accuracy on the two datasets, respectively.
comment: 13 pages. This paper is accepted by ICSE 2025
☆ ContextModule: Improving Code Completion via Repository-level Contextual Information
Large Language Models (LLMs) have demonstrated impressive capabilities in code completion tasks, where they assist developers by predicting and generating new code in real-time. However, existing LLM-based code completion systems primarily rely on the immediate context of the file being edited, often missing valuable repository-level information, user behaviour and edit history that could improve suggestion accuracy. Additionally, challenges such as efficiently retrieving relevant code snippets from large repositories, incorporating user behavior, and balancing accuracy with low-latency requirements in production environments remain unresolved. In this paper, we propose ContextModule, a framework designed to enhance LLM-based code completion by retrieving and integrating three types of contextual information from the repository: user behavior-based code, similar code snippets, and critical symbol definitions. By capturing user interactions across files and leveraging repository-wide static analysis, ContextModule improves the relevance and precision of generated code. We implement performance optimizations, such as index caching, to ensure the system meets the latency constraints of real-world coding environments. Experimental results and industrial practise demonstrate that ContextModule significantly improves code completion accuracy and user acceptance rates.
☆ Go-Oracle: Automated Test Oracle for Go Concurrency Bugs
The Go programming language has gained significant traction for developing software, especially in various infrastructure systems. Nonetheless, concurrency bugs have become a prevalent issue within Go, presenting a unique challenge due to the language's dual concurrency mechanisms-communicating sequential processes and shared memory. Detecting concurrency bugs and accurately classifying program executions as pass or fail presents an immense challenge, even for domain experts. We conducted a survey with expert developers at Bytedance that confirmed this challenge. Our work seeks to address the test oracle problem for Go programs, to automatically classify test executions as pass or fail. This problem has not been investigated in the literature for Go programs owing to its distinctive programming model. Our approach involves collecting both passing and failing execution traces from various subject Go programs. We capture a comprehensive array of execution events using the native Go execution tracer. Subsequently, we preprocess and encode these traces before training a transformer-based neural network to effectively classify the traces as either passing or failing. The evaluation of our approach encompasses 8 subject programs sourced from the GoBench repository. These subject programs are routinely used as benchmarks in an industry setting. Encouragingly, our test oracle, Go-Oracle, demonstrates high accuracies even when operating with a limited dataset, showcasing the efficacy and potential of our methodology. Developers at Bytedance strongly agreed that they would use the Go-Oracle tool over the current practice of manual inspections to classify tests for Go programs as pass or fail.
☆ Federated In-Context LLM Agent Learning
Large Language Models (LLMs) have revolutionized intelligent services by enabling logical reasoning, tool use, and interaction with external systems as agents. The advancement of LLMs is frequently hindered by the scarcity of high-quality data, much of which is inherently sensitive. Federated learning (FL) offers a potential solution by facilitating the collaborative training of distributed LLMs while safeguarding private data. However, FL frameworks face significant bandwidth and computational demands, along with challenges from heterogeneous data distributions. The emerging in-context learning capability of LLMs offers a promising approach by aggregating natural language rather than bulky model parameters. Yet, this method risks privacy leakage, as it necessitates the collection and presentation of data samples from various clients during aggregation. In this paper, we propose a novel privacy-preserving Federated In-Context LLM Agent Learning (FICAL) algorithm, which to our best knowledge for the first work unleashes the power of in-context learning to train diverse LLM agents through FL. In our design, knowledge compendiums generated by a novel LLM-enhanced Knowledge Compendiums Generation (KCG) module are transmitted between clients and the server instead of model parameters in previous FL methods. Apart from that, an incredible Retrieval Augmented Generation (RAG) based Tool Learning and Utilizing (TLU) module is designed and we incorporate the aggregated global knowledge compendium as a teacher to teach LLM agents the usage of tools. We conducted extensive experiments and the results show that FICAL has competitive performance compared to other SOTA baselines with a significant communication cost decrease of $\mathbf{3.33\times10^5}$ times.
☆ DynamicPAE: Generating Scene-Aware Physical Adversarial Examples in Real-Time
Physical adversarial examples (PAEs) are regarded as "whistle-blowers" of real-world risks in deep-learning applications. However, current PAE generation studies show limited adaptive attacking ability to diverse and varying scenes. The key challenges in generating dynamic PAEs are exploring their patterns under noisy gradient feedback and adapting the attack to agnostic scenario natures. To address the problems, we present DynamicPAE, the first generative framework that enables scene-aware real-time physical attacks beyond static attacks. Specifically, to train the dynamic PAE generator under noisy gradient feedback, we introduce the residual-driven sample trajectory guidance technique, which redefines the training task to break the limited feedback information restriction that leads to the degeneracy problem. Intuitively, it allows the gradient feedback to be passed to the generator through a low-noise auxiliary task, thereby guiding the optimization away from degenerate solutions and facilitating a more comprehensive and stable exploration of feasible PAEs. To adapt the generator to agnostic scenario natures, we introduce the context-aligned scene expectation simulation process, consisting of the conditional-uncertainty-aligned data module and the skewness-aligned objective re-weighting module. The former enhances robustness in the context of incomplete observation by employing a conditional probabilistic model for domain randomization, while the latter facilitates consistent stealth control across different attack targets by automatically reweighting losses based on the skewness indicator. Extensive digital and physical evaluations demonstrate the superior attack performance of DynamicPAE, attaining a 1.95 $\times$ boost (65.55% average AP drop under attack) on representative object detectors (e.g., Yolo-v8) over state-of-the-art static PAE generating methods.
comment: This work has been submitted to the IEEE for possible publication
☆ NeRF-NQA: No-Reference Quality Assessment for Scenes Generated by NeRF and Neural View Synthesis Methods
Neural View Synthesis (NVS) has demonstrated efficacy in generating high-fidelity dense viewpoint videos using a image set with sparse views. However, existing quality assessment methods like PSNR, SSIM, and LPIPS are not tailored for the scenes with dense viewpoints synthesized by NVS and NeRF variants, thus, they often fall short in capturing the perceptual quality, including spatial and angular aspects of NVS-synthesized scenes. Furthermore, the lack of dense ground truth views makes the full reference quality assessment on NVS-synthesized scenes challenging. For instance, datasets such as LLFF provide only sparse images, insufficient for complete full-reference assessments. To address the issues above, we propose NeRF-NQA, the first no-reference quality assessment method for densely-observed scenes synthesized from the NVS and NeRF variants. NeRF-NQA employs a joint quality assessment strategy, integrating both viewwise and pointwise approaches, to evaluate the quality of NVS-generated scenes. The viewwise approach assesses the spatial quality of each individual synthesized view and the overall inter-views consistency, while the pointwise approach focuses on the angular qualities of scene surface points and their compound inter-point quality. Extensive evaluations are conducted to compare NeRF-NQA with 23 mainstream visual quality assessment methods (from fields of image, video, and light-field assessment). The results demonstrate NeRF-NQA outperforms the existing assessment methods significantly and it shows substantial superiority on assessing NVS-synthesized scenes without references. An implementation of this paper are available at https://github.com/VincentQQu/NeRF-NQA.
☆ Can a MISL Fly? Analysis and Ingredients for Mutual Information Skill Learning
Self-supervised learning has the potential of lifting several of the key challenges in reinforcement learning today, such as exploration, representation learning, and reward design. Recent work (METRA) has effectively argued that moving away from mutual information and instead optimizing a certain Wasserstein distance is important for good performance. In this paper, we argue that the benefits seen in that paper can largely be explained within the existing framework of mutual information skill learning (MISL). Our analysis suggests a new MISL method (contrastive successor features) that retains the excellent performance of METRA with fewer moving parts, and highlights connections between skill learning, contrastive representation learning, and successor features. Finally, through careful ablation studies, we provide further insight into some of the key ingredients for both our method and METRA.
comment: Code and videos are available on the website: https://princeton-rl.github.io/contrastive-successor-features/
☆ Intelligent Control of Robotic X-ray Devices using a Language-promptable Digital Twin
Natural language offers a convenient, flexible interface for controlling robotic C-arm X-ray systems, making advanced functionality and controls accessible. However, enabling language interfaces requires specialized AI models that interpret X-ray images to create a semantic representation for reasoning. The fixed outputs of such AI models limit the functionality of language controls. Incorporating flexible, language-aligned AI models prompted through language enables more versatile interfaces for diverse tasks and procedures. Using a language-aligned foundation model for X-ray image segmentation, our system continually updates a patient digital twin based on sparse reconstructions of desired anatomical structures. This supports autonomous capabilities such as visualization, patient-specific viewfinding, and automatic collimation from novel viewpoints, enabling commands 'Focus in on the lower lumbar vertebrae.' In a cadaver study, users visualized, localized, and collimated structures across the torso using verbal commands, achieving 84% end-to-end success. Post hoc analysis of randomly oriented images showed our patient digital twin could localize 35 commonly requested structures to within 51.68 mm, enabling localization and isolation from arbitrary orientations. Our results demonstrate how intelligent robotic X-ray systems can incorporate physicians' expressed intent directly. While existing foundation models for intra-operative X-ray analysis exhibit failure modes, as they improve, they can facilitate highly flexible, intelligent robotic C-arms.
☆ MAGIC: Mastering Physical Adversarial Generation in Context through Collaborative LLM Agents
Physical adversarial attacks in driving scenarios can expose critical vulnerabilities in visual perception models. However, developing such attacks remains challenging due to diverse real-world backgrounds and the requirement for maintaining visual naturality. Building upon this challenge, we reformulate physical adversarial attacks as a one-shot patch-generation problem. Our approach generates adversarial patches through a deep generative model that considers the specific scene context, enabling direct physical deployment in matching environments. The primary challenge lies in simultaneously achieving two objectives: generating adversarial patches that effectively mislead object detection systems while determining contextually appropriate placement within the scene. We propose MAGIC (Mastering Physical Adversarial Generation In Context), a novel framework powered by multi-modal LLM agents to address these challenges. MAGIC automatically understands scene context and orchestrates adversarial patch generation through the synergistic interaction of language and vision capabilities. MAGIC orchestrates three specialized LLM agents: The adv-patch generation agent (GAgent) masters the creation of deceptive patches through strategic prompt engineering for text-to-image models. The adv-patch deployment agent (DAgent) ensures contextual coherence by determining optimal placement strategies based on scene understanding. The self-examination agent (EAgent) completes this trilogy by providing critical oversight and iterative refinement of both processes. We validate our method on both digital and physical level, \ie, nuImage and manually captured real scenes, where both statistical and visual results prove that our MAGIC is powerful and effectively for attacking wide-used object detection systems.
☆ Adaptive Querying for Reward Learning from Human Feedback
Learning from human feedback is a popular approach to train robots to adapt to user preferences and improve safety. Existing approaches typically consider a single querying (interaction) format when seeking human feedback and do not leverage multiple modes of user interaction with a robot. We examine how to learn a penalty function associated with unsafe behaviors, such as side effects, using multiple forms of human feedback, by optimizing the query state and feedback format. Our framework for adaptive feedback selection enables querying for feedback in critical states in the most informative format, while accounting for the cost and probability of receiving feedback in a certain format. We employ an iterative, two-phase approach which first selects critical states for querying, and then uses information gain to select a feedback format for querying across the sampled critical states. Our evaluation in simulation demonstrates the sample efficiency of our approach.
♻ ☆ Length Optimization in Conformal Prediction
Conditional validity and length efficiency are two crucial aspects of conformal prediction (CP). Conditional validity ensures accurate uncertainty quantification for data subpopulations, while proper length efficiency ensures that the prediction sets remain informative. Despite significant efforts to address each of these issues individually, a principled framework that reconciles these two objectives has been missing in the CP literature. In this paper, we develop Conformal Prediction with Length-Optimization (CPL) - a novel and practical framework that constructs prediction sets with (near-) optimal length while ensuring conditional validity under various classes of covariate shifts, including the key cases of marginal and group-conditional coverage. In the infinite sample regime, we provide strong duality results which indicate that CPL achieves conditional validity and length optimality. In the finite sample regime, we show that CPL constructs conditionally valid prediction sets. Our extensive empirical evaluations demonstrate the superior prediction set size performance of CPL compared to state-of-the-art methods across diverse real-world and synthetic datasets in classification, regression, and large language model-based multiple choice question answering. An Implementation of our algorithm can be accessed at the following link: https://github.com/shayankiyani98/CP.
♻ ☆ Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Change KDD
Counterfactual explanations (CFEs) guide users on how to adjust inputs to machine learning models to achieve desired outputs. While existing research primarily addresses static scenarios, real-world applications often involve data or model changes, potentially invalidating previously generated CFEs and rendering user-induced input changes ineffective. Current methods addressing this issue often support only specific models or change types, require extensive hyperparameter tuning, or fail to provide probabilistic guarantees on CFE robustness to model changes. This paper proposes a novel approach for generating CFEs that provides probabilistic guarantees for any model and change type, while offering interpretable and easy-to-select hyperparameters. We establish a theoretical framework for probabilistically defining robustness to model change and demonstrate how our BetaRCE method directly stems from it. BetaRCE is a post-hoc method applied alongside a chosen base CFE generation method to enhance the quality of the explanation beyond robustness. It facilitates a transition from the base explanation to a more robust one with user-adjusted probability bounds. Through experimental comparisons with baselines, we show that BetaRCE yields robust, most plausible, and closest to baseline counterfactual explanations.
comment: Accepted at 31st SIGKDD Conference on Knowledge Discovery and Data Mining 2025
♻ ☆ ProVision: Programmatically Scaling Vision-centric Instruction Data for Multimodal Language Models
With the rise of multimodal applications, instruction data has become critical for training multimodal language models capable of understanding complex image-based queries. Existing practices rely on powerful but costly large language models (LLMs) or multimodal language models (MLMs) to produce instruction data. These are often prone to hallucinations, licensing issues and the generation process is often hard to scale and interpret. In this work, we present a programmatic approach that employs scene graphs as symbolic representations of images and human-written programs to systematically synthesize vision-centric instruction data. Our approach ensures the interpretability and controllability of the data generation process and scales efficiently while maintaining factual accuracy. By implementing a suite of 24 single-image, 14 multi-image instruction generators, and a scene graph generation pipeline, we build a scalable, cost-effective system: ProVision which produces diverse question-answer pairs concerning objects, attributes, relations, depth, etc., for any given image. Applied to Visual Genome and DataComp datasets, we generate over 10 million instruction data points, ProVision-10M, and leverage them in both pretraining and instruction tuning stages of MLMs. When adopted in the instruction tuning stage, our single-image instruction data yields up to a 7% improvement on the 2D split and 8% on the 3D split of CVBench, along with a 3% increase in performance on QBench2, RealWorldQA, and MMMU. Our multi-image instruction data leads to an 8% improvement on Mantis-Eval. Incorporation of our data in both pre-training and fine-tuning stages of xGen-MM-4B leads to an averaged improvement of 1.6% across 11 benchmarks.
comment: code: https://github.com/JieyuZ2/ProVision dataset: https://huggingface.co/datasets/Salesforce/ProVision-10M
♻ ☆ Fusing Domain-Specific Content from Large Language Models into Knowledge Graphs for Enhanced Zero Shot Object State Classification AAAI
Domain-specific knowledge can significantly contribute to addressing a wide variety of vision tasks. However, the generation of such knowledge entails considerable human labor and time costs. This study investigates the potential of Large Language Models (LLMs) in generating and providing domain-specific information through semantic embeddings. To achieve this, an LLM is integrated into a pipeline that utilizes Knowledge Graphs and pre-trained semantic vectors in the context of the Vision-based Zero-shot Object State Classification task. We thoroughly examine the behavior of the LLM through an extensive ablation study. Our findings reveal that the integration of LLM-based embeddings, in combination with general-purpose pre-trained embeddings, leads to substantial performance improvements. Drawing insights from this ablation study, we conduct a comparative analysis against competing models, thereby highlighting the state-of-the-art performance achieved by the proposed approach.
comment: Accepted at the AAAI-MAKE 2024
♻ ☆ Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Prompts
As large language models (LLMs) become increasingly prevalent across many real-world applications, understanding and enhancing their robustness to adversarial attacks is of paramount importance. Existing methods for identifying adversarial prompts tend to focus on specific domains, lack diversity, or require extensive human annotations. To address these limitations, we present Rainbow Teaming, a novel black-box approach for producing a diverse collection of adversarial prompts. Rainbow Teaming casts adversarial prompt generation as a quality-diversity problem and uses open-ended search to generate prompts that are both effective and diverse. Focusing on the safety domain, we use Rainbow Teaming to target various state-of-the-art LLMs, including the Llama 2 and Llama 3 models. Our approach reveals hundreds of effective adversarial prompts, with an attack success rate exceeding 90% across all tested models. Furthermore, we demonstrate that prompts generated by Rainbow Teaming are highly transferable and that fine-tuning models with synthetic data generated by our method significantly enhances their safety without sacrificing general performance or helpfulness. We additionally explore the versatility of Rainbow Teaming by applying it to question answering and cybersecurity, showcasing its potential to drive robust open-ended self-improvement in a wide range of applications.
♻ ☆ The BrowserGym Ecosystem for Web Agent Research
The BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents, particularly those leveraging automation and Large Language Models (LLMs) for web interaction tasks. Many existing benchmarks suffer from fragmentation and inconsistent evaluation methodologies, making it challenging to achieve reliable comparisons and reproducible results. BrowserGym aims to solve this by providing a unified, gym-like environment with well-defined observation and action spaces, facilitating standardized evaluation across diverse benchmarks. Combined with AgentLab, a complementary framework that aids in agent creation, testing, and analysis, BrowserGym offers flexibility for integrating new benchmarks while ensuring consistent evaluation and comprehensive experiment management. This standardized approach seeks to reduce the time and complexity of developing web agents, supporting more reliable comparisons and facilitating in-depth analysis of agent behaviors, and could result in more adaptable, capable agents, ultimately accelerating innovation in LLM-driven automation. As a supporting evidence, we conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across all benchmarks currently available in BrowserGym. Among other findings, our results highlight a large discrepancy between OpenAI and Anthropic's latests models, with Claude-3.5-Sonnet leading the way on almost all benchmarks, except on vision-related tasks where GPT-4o is superior. Despite these advancements, our results emphasize that building robust and efficient web agents remains a significant challenge, due to the inherent complexity of real-world web environments and the limitations of current models.
♻ ☆ Comparative Analysis of Pooling Mechanisms in LLMs: A Sentiment Analysis Perspective
Large Language Models (LLMs) have revolutionized natural language processing (NLP) by delivering state-of-the-art performance across a variety of tasks. Among these, Transformer-based models like BERT and GPT rely on pooling layers to aggregate token-level embeddings into sentence-level representations. Common pooling mechanisms such as Mean, Max, and Weighted Sum play a pivotal role in this aggregation process. Despite their widespread use, the comparative performance of these strategies on different LLM architectures remains underexplored. To address this gap, this paper investigates the effects of these pooling mechanisms on two prominent LLM families -- BERT and GPT, in the context of sentence-level sentiment analysis. Comprehensive experiments reveal that each pooling mechanism exhibits unique strengths and weaknesses depending on the task's specific requirements. Our findings underline the importance of selecting pooling methods tailored to the demands of particular applications, prompting a re-evaluation of common assumptions regarding pooling operations. By offering actionable insights, this study contributes to the optimization of LLM-based models for downstream tasks.
comment: 4 figures
♻ ☆ Weighted Ensemble Models Are Strong Continual Learners ECCV2024
In this work, we study the problem of continual learning (CL) where the goal is to learn a model on a sequence of tasks, such that the data from the previous tasks becomes unavailable while learning on the current task data. CL is essentially a balancing act between being able to learn on the new task (i.e., plasticity) and maintaining the performance on the previously learned concepts (i.e., stability). Intending to address the stability-plasticity trade-off, we propose to perform weight-ensembling of the model parameters of the previous and current tasks. This weighted-ensembled model, which we call Continual Model Averaging (or CoMA), attains high accuracy on the current task by leveraging plasticity, while not deviating too far from the previous weight configuration, ensuring stability. We also propose an improved variant of CoMA, named Continual Fisher-weighted Model Averaging (or CoFiMA), that selectively weighs each parameter in the weights ensemble by leveraging the Fisher information of the weights of the model. Both variants are conceptually simple, easy to implement, and effective in attaining state-of-the-art performance on several standard CL benchmarks. Code is available at: https://github.com/IemProg/CoFiMA.
comment: Accepted for ECCV2024, Code: https://github.com/IemProg/CoFiMA
♻ ☆ AI Can Enhance Creativity in Social Networks
Can peer recommendation engines elevate people's creative performances in self-organizing social networks? Answering this question requires resolving challenges in data collection (e.g., tracing inspiration links and psycho-social attributes of nodes) and intervention design (e.g., balancing idea stimulation and redundancy in evolving information environments). We trained a model that predicts people's ideation performances using semantic and network-structural features in an online platform. Using this model, we built SocialMuse, which maximizes people's predicted performances to generate peer recommendations for them. We found treatment networks leveraging SocialMuse outperforming AI-agnostic control networks in several creativity measures. The treatment networks were more decentralized than the control, as SocialMuse increasingly emphasized network-structural features at large network sizes. This decentralization spreads people's inspiration sources, helping inspired ideas stand out better. Our study provides actionable insights into building intelligent systems for elevating creativity.
♻ ☆ AgentMixer: Multi-Agent Correlated Policy Factorization
In multi-agent reinforcement learning, centralized training with decentralized execution (CTDE) methods typically assume that agents make decisions based on their local observations independently, which may not lead to a correlated joint policy with coordination. Coordination can be explicitly encouraged during training and individual policies can be trained to imitate the correlated joint policy. However, this may lead to an \textit{asymmetric learning failure} due to the observation mismatch between the joint and individual policies. Inspired by the concept of correlated equilibrium, we introduce a \textit{strategy modification} called AgentMixer that allows agents to correlate their policies. AgentMixer combines individual partially observable policies into a joint fully observable policy non-linearly. To enable decentralized execution, we introduce \textit{Individual-Global-Consistency} to guarantee mode consistency during joint training of the centralized and decentralized policies and prove that AgentMixer converges to an $\epsilon$-approximate Correlated Equilibrium. In the Multi-Agent MuJoCo, SMAC-v2, Matrix Game, and Predator-Prey benchmarks, AgentMixer outperforms or matches state-of-the-art methods.
♻ ☆ Extreme AutoML: Analysis of Classification, Regression, and NLP Performance
Utilizing machine learning techniques has always required choosing hyperparameters. This is true whether one uses a classical technique such as a KNN or very modern neural networks such as Deep Learning. Though in many applications, hyperparameters are chosen by hand, automated methods have become increasingly more common. These automated methods have become collectively known as automated machine learning, or AutoML. Several automated selection algorithms have shown similar or improved performance over state-of-the-art methods. This breakthrough has led to the development of cloud-based services like Google AutoML, which is based on Deep Learning and is widely considered to be the industry leader in AutoML services. Extreme Learning Machines (ELMs) use a fundamentally different type of neural architecture, producing better results at a significantly discounted computational cost. We benchmark the Extreme AutoML technology against Google's AutoML using several popular classification data sets from the University of California at Irvine's (UCI) repository, and several other data sets, observing significant advantages for Extreme AutoML in accuracy, Jaccard Indices, the variance of Jaccard Indices across classes (i.e. class variance) and training times.
comment: 8 pages, 7 figures
♻ ☆ VoiceBench: Benchmarking LLM-Based Voice Assistants
Building on the success of large language models (LLMs), recent advancements such as GPT-4o have enabled real-time speech interactions through LLM-based voice assistants, offering a significantly improved user experience compared to traditional text-based interactions. However, the absence of benchmarks designed to evaluate these speech interaction capabilities has hindered progress of LLM-based voice assistants development. Current evaluations focus primarily on automatic speech recognition (ASR) or general knowledge evaluation with clean speeches, neglecting the more intricate, real-world scenarios that involve diverse speaker characteristics, environmental and content factors. To address this, we introduce VoiceBench, the first benchmark designed to provide a multi-faceted evaluation of LLM-based voice assistants. VoiceBench also includes both real and synthetic spoken instructions that incorporate the above three key real-world variations. Extensive experiments reveal the limitations of current LLM-based voice assistant models and offer valuable insights for future research and development in this field.
comment: Work in progress. Data is available at https://github.com/MatthewCYM/VoiceBench
♻ ☆ Topic Classification of Case Law Using a Large Language Model and a New Taxonomy for UK Law: AI Insights into Summary Judgment
This paper addresses a critical gap in legal analytics by developing and applying a novel taxonomy for topic classification of summary judgment cases in the United Kingdom. Using a curated dataset of summary judgment cases, we use the Large Language Model Claude 3 Opus to explore functional topics and trends. We find that Claude 3 Opus correctly classified the topic with an accuracy of 87.13% and an F1 score of 0.87. The analysis reveals distinct patterns in the application of summary judgments across various legal domains. As case law in the United Kingdom is not originally labelled with keywords or a topic filtering option, the findings not only refine our understanding of the thematic underpinnings of summary judgments but also illustrate the potential of combining traditional and AI-driven approaches in legal classification. Therefore, this paper provides a new and general taxonomy for UK law. The implications of this work serve as a foundation for further research and policy discussions in the field of judicial administration and computational legal research methodologies.
♻ ☆ Shaping AI's Impact on Billions of Lives
Artificial Intelligence (AI), like any transformative technology, has the potential to be a double-edged sword, leading either toward significant advancements or detrimental outcomes for society as a whole. As is often the case when it comes to widely-used technologies in market economies (e.g., cars and semiconductor chips), commercial interest tends to be the predominant guiding factor. The AI community is at risk of becoming polarized to either take a laissez-faire attitude toward AI development, or to call for government overregulation. Between these two poles we argue for the community of AI practitioners to consciously and proactively work for the common good. This paper offers a blueprint for a new type of innovation infrastructure including 18 concrete milestones to guide AI research in that direction. Our view is that we are still in the early days of practical AI, and focused efforts by practitioners, policymakers, and other stakeholders can still maximize the upsides of AI and minimize its downsides. We talked to luminaries such as recent Nobelist John Jumper on science, President Barack Obama on governance, former UN Ambassador and former National Security Advisor Susan Rice on security, philanthropist Eric Schmidt on several topics, and science fiction novelist Neal Stephenson on entertainment. This ongoing dialogue and collaborative effort has produced a comprehensive, realistic view of what the actual impact of AI could be, from a diverse assembly of thinkers with deep understanding of this technology and these domains. From these exchanges, five recurring guidelines emerged, which form the cornerstone of a framework for beginning to harness AI in service of the public good. They not only guide our efforts in discovery but also shape our approach to deploying this transformative technology responsibly and ethically.
♻ ☆ Annotation-guided Protein Design with Multi-Level Domain Alignment KDD 2025
The core challenge of de novo protein design lies in creating proteins with specific functions or properties, guided by certain conditions. Current models explore to generate protein using structural and evolutionary guidance, which only provide indirect conditions concerning functions and properties. However, textual annotations of proteins, especially the annotations for protein domains, which directly describe the protein's high-level functionalities, properties, and their correlation with target amino acid sequences, remain unexplored in the context of protein design tasks. In this paper, we propose Protein-Annotation Alignment Generation, PAAG, a multi-modality protein design framework that integrates the textual annotations extracted from protein database for controllable generation in sequence space. Specifically, within a multi-level alignment module, PAAG can explicitly generate proteins containing specific domains conditioned on the corresponding domain annotations, and can even design novel proteins with flexible combinations of different kinds of annotations. Our experimental results underscore the superiority of the aligned protein representations from PAAG over 7 prediction tasks. Furthermore, PAAG demonstrates a significant increase in generation success rate (24.7% vs 4.7% in zinc finger, and 54.3% vs 22.0% in the immunoglobulin domain) in comparison to the existing model. We anticipate that PAAG will broaden the horizons of protein design by leveraging the knowledge from between textual annotation and proteins.
comment: Accepted by KDD 2025
♻ ☆ EmoSpeech: A Corpus of Emotionally Rich and Contextually Detailed Speech Annotations
Advances in text-to-speech (TTS) technology have significantly improved the quality of generated speech, closely matching the timbre and intonation of the target speaker. However, due to the inherent complexity of human emotional expression, the development of TTS systems capable of controlling subtle emotional differences remains a formidable challenge. Existing emotional speech databases often suffer from overly simplistic labelling schemes that fail to capture a wide range of emotional states, thus limiting the effectiveness of emotion synthesis in TTS applications. To this end, recent efforts have focussed on building databases that use natural language annotations to describe speech emotions. However, these approaches are costly and require more emotional depth to train robust systems. In this paper, we propose a novel process aimed at building databases by systematically extracting emotion-rich speech segments and annotating them with detailed natural language descriptions through a generative model. This approach enhances the emotional granularity of the database and significantly reduces the reliance on costly manual annotations by automatically augmenting the data with high-level language models. The resulting rich database provides a scalable and economically viable solution for developing a more nuanced and dynamic basis for developing emotionally controlled TTS systems.
comment: We understand that this article is not currently allowed to be published, and there are major errors in the data, so it needs to be withdrawn and revised
♻ ☆ Local vs. Global: Local Land-Use and Land-Cover Models Deliver Higher Quality Maps
In 2023, 58.0% of the African population experienced moderate to severe food insecurity, with 21.6% facing severe food insecurity. Land-use and land-cover maps provide crucial insights for addressing food insecurity by improving agricultural efforts, including mapping and monitoring crop types and estimating yield. The development of global land-cover maps has been facilitated by the increasing availability of earth observation data and advancements in geospatial machine learning. However, these global maps exhibit lower accuracy and inconsistencies in Africa, partly due to the lack of representative training data. To address this issue, we propose a data-centric framework with a teacher-student model setup, which uses diverse data sources of satellite images and label examples to produce local land-cover maps. Our method trains a high-resolution teacher model on images with a resolution of 0.331 m/pixel and a low-resolution student model on publicly available images with a resolution of 10 m/pixel. The student model also utilizes the teacher model's output as its weak label examples through knowledge transfer. We evaluated our framework using Murang'a county in Kenya, renowned for its agricultural productivity, as a use case. Our local models achieved higher quality maps, with improvements of 0.14 in the F1 score and 0.21 in Intersection-over-Union, compared to the best global model. Our evaluation also revealed inconsistencies in existing global maps, with a maximum agreement rate of 0.30 among themselves. Our work provides valuable guidance to decision-makers for driving informed decisions to enhance food security.
♻ ☆ Robust Deep Reinforcement Learning Through Adversarial Attacks and Training : A Survey
Deep Reinforcement Learning (DRL) is a subfield of machine learning for training autonomous agents that take sequential actions across complex environments. Despite its significant performance in well-known environments, it remains susceptible to minor condition variations, raising concerns about its reliability in real-world applications. To improve usability, DRL must demonstrate trustworthiness and robustness. A way to improve the robustness of DRL to unknown changes in the environmental conditions and possible perturbations is through Adversarial Training, by training the agent against well-suited adversarial attacks on the observations and the dynamics of the environment. Addressing this critical issue, our work presents an in-depth analysis of contemporary adversarial attack and training methodologies, systematically categorizing them and comparing their objectives and operational mechanisms.
comment: 61 pages, 17 figues, 1 table
♻ ☆ What Should We Engineer in Prompts? Training Humans in Requirement-Driven LLM Use
Prompting LLMs for complex tasks (e.g., building a trip advisor chatbot) needs humans to clearly articulate customized requirements (e.g., "start the response with a tl;dr"). However, existing prompt engineering instructions often lack focused training on requirement articulation and instead tend to emphasize increasingly automatable strategies (e.g., tricks like adding role-plays and "think step-by-step"). To address the gap, we introduce Requirement-Oriented Prompt Engineering (ROPE), a paradigm that focuses human attention on generating clear, complete requirements during prompting. We implement ROPE through an assessment and training suite that provides deliberate practice with LLM-generated feedback. In a randomized controlled experiment with 30 novices, ROPE significantly outperforms conventional prompt engineering training (20% vs. 1% gains), a gap that automatic prompt optimization cannot close. Furthermore, we demonstrate a direct correlation between the quality of input requirements and LLM outputs. Our work paves the way to empower more end-users to build complex LLM applications.
comment: 15 pages
♻ ☆ Spatial-Temporal Federated Learning for Lifelong Person Re-identification on Distributed Edges
Data drift is a thorny challenge when deploying person re-identification (ReID) models into real-world devices, where the data distribution is significantly different from that of the training environment and keeps changing. To tackle this issue, we propose a federated spatial-temporal incremental learning approach, named FedSTIL, which leverages both lifelong learning and federated learning to continuously optimize models deployed on many distributed edge clients. Unlike previous efforts, FedSTIL aims to mine spatial-temporal correlations among the knowledge learnt from different edge clients. Specifically, the edge clients first periodically extract general representations of drifted data to optimize their local models. Then, the learnt knowledge from edge clients will be aggregated by centralized parameter server, where the knowledge will be selectively and attentively distilled from spatial- and temporal-dimension with carefully designed mechanisms. Finally, the distilled informative spatial-temporal knowledge will be sent back to correlated edge clients to further improve the recognition accuracy of each edge client with a lifelong learning method. Extensive experiments on a mixture of five real-world datasets demonstrate that our method outperforms others by nearly 4% in Rank-1 accuracy, while reducing communication cost by 62%. All implementation codes are publicly available on https://github.com/MSNLAB/Federated-Lifelong-Person-ReID
♻ ☆ Social Recommendation through Heterogeneous Graph Modeling of the Long-term and Short-term Preference Defined by Dynamic Time Spans
Social recommendations have been widely adopted in substantial domains. Recently, graph neural networks (GNN) have been employed in recommender systems due to their success in graph representation learning. However, dealing with the dynamic property of social network data is a challenge. This research presents a novel method that provides social recommendations by incorporating the dynamic property of social network data in a heterogeneous graph. The model aims to capture user preference over time without going through the complexities of a dynamic graph by adding period nodes to define users' long-term and short-term preferences and aggregating assigned edge weights. The model is applied to real-world data to argue its superior performance. Promising results demonstrate the effectiveness of this model.
♻ ☆ Graph Agent Network: Empowering Nodes with Inference Capabilities for Adversarial Resilience
End-to-end training with global optimization have popularized graph neural networks (GNNs) for node classification, yet inadvertently introduced vulnerabilities to adversarial edge-perturbing attacks. Adversaries can exploit the inherent opened interfaces of GNNs' input and output, perturbing critical edges and thus manipulating the classification results. Current defenses, due to their persistent utilization of global-optimization-based end-to-end training schemes, inherently encapsulate the vulnerabilities of GNNs. This is specifically evidenced in their inability to defend against targeted secondary attacks. In this paper, we propose the Graph Agent Network (GAgN) to address the aforementioned vulnerabilities of GNNs. GAgN is a graph-structured agent network in which each node is designed as an 1-hop-view agent. Through the decentralized interactions between agents, they can learn to infer global perceptions to perform tasks including inferring embeddings, degrees and neighbor relationships for given nodes. This empowers nodes to filtering adversarial edges while carrying out classification tasks. Furthermore, agents' limited view prevents malicious messages from propagating globally in GAgN, thereby resisting global-optimization-based secondary attacks. We prove that single-hidden-layer multilayer perceptrons (MLPs) are theoretically sufficient to achieve these functionalities. Experimental results show that GAgN effectively implements all its intended capabilities and, compared to state-of-the-art defenses, achieves optimal classification accuracy on the perturbed datasets.
♻ ☆ Diffusion Model Patching via Mixture-of-Prompts AAAI 2025
We present Diffusion Model Patching (DMP), a simple method to boost the performance of pre-trained diffusion models that have already reached convergence, with a negligible increase in parameters. DMP inserts a small, learnable set of prompts into the model's input space while keeping the original model frozen. The effectiveness of DMP is not merely due to the addition of parameters but stems from its dynamic gating mechanism, which selects and combines a subset of learnable prompts at every timestep (i.e., reverse denoising steps). This strategy, which we term "mixture-of-prompts", enables the model to draw on the distinct expertise of each prompt, essentially "patching" the model's functionality at every timestep with minimal yet specialized parameters. Uniquely, DMP enhances the model by further training on the original dataset already used for pre-training, even in a scenario where significant improvements are typically not expected due to model convergence. Notably, DMP significantly enhances the FID of converged DiT-L/2 by 10.38% on FFHQ, achieved with only a 1.43% parameter increase and 50K additional training iterations.
comment: AAAI 2025; Project: https://sangminwoo.github.io/DMP/
♻ ☆ GEE-OPs: An Operator Knowledge Base for Geospatial Code Generation on the Google Earth Engine Platform Powered by Large Language Models
As the scale and complexity of spatiotemporal data continue to grow rapidly, the use of geospatial modeling on the Google Earth Engine (GEE) platform presents dual challenges: improving the coding efficiency of domain experts and enhancing the coding capabilities of interdisciplinary users. To address these challenges and improve the performance of large language models (LLMs) in geospatial code generation tasks, we propose a framework for building a geospatial operator knowledge base tailored to the GEE JavaScript API. This framework consists of an operator syntax knowledge table, an operator relationship frequency table, an operator frequent pattern knowledge table, and an operator relationship chain knowledge table. By leveraging Abstract Syntax Tree (AST) techniques and frequent itemset mining, we systematically extract operator knowledge from 185,236 real GEE scripts and syntax documentation, forming a structured knowledge base. Experimental results demonstrate that the framework achieves over 90% accuracy, recall, and F1 score in operator knowledge extraction. When integrated with the Retrieval-Augmented Generation (RAG) strategy for LLM-based geospatial code generation tasks, the knowledge base improves performance by 20-30%. Ablation studies further quantify the necessity of each knowledge table in the knowledge base construction. This work provides robust support for the advancement and application of geospatial code modeling techniques, offering an innovative approach to constructing domain-specific knowledge bases that enhance the code generation capabilities of LLMs, and fostering the deeper integration of generative AI technologies within the field of geoinformatics.
♻ ☆ Flash Communication: Reducing Tensor Parallelization Bottleneck for Fast Large Language Model Inference
The ever-increasing sizes of large language models necessitate distributed solutions for fast inference that exploit multi-dimensional parallelism, where computational loads are split across various accelerators such as GPU clusters. However, this approach often introduces significant communication overhead, especially on devices with limited bandwidth. In this paper, we introduce Flash Communication, a novel low-bit compression technique designed to alleviate the tensor-parallelism communication bottleneck during inference. Our method substantially boosts intra-node communication speed by more than 3x and reduces the time-to-first-token by 2x, with nearly no sacrifice in model accuracy. Extensive experiments on various up-to-date LLMs demonstrate the effectiveness of our approach.
♻ ☆ Understanding Token Probability Encoding in Output Embeddings COLING 2025
In this paper, we investigate the output token probability information in the output embedding of language models. We find an approximate common log-linear encoding of output token probabilities within the output embedding vectors and empirically demonstrate that it is accurate and sparse. As a causality examination, we steer the encoding in output embedding to modify the output probability distribution accurately. Moreover, the sparsity we find in output probability encoding suggests that a large number of dimensions in the output embedding do not contribute to causal language modeling. Therefore, we attempt to delete the output-unrelated dimensions and find more than 30% of the dimensions can be deleted without significant movement in output distribution and sequence generation. Additionally, in the pre-training dynamics of language models, we find that the output embeddings capture the corpus token frequency information in early steps, even before an obvious convergence of parameters starts.
comment: 15 pages, 17 figures, 3 tables. COLING 2025 Accepted
♻ ☆ M2OST: Many-to-one Regression for Predicting Spatial Transcriptomics from Digital Pathology Images AAAI 2025
The advancement of Spatial Transcriptomics (ST) has facilitated the spatially-aware profiling of gene expressions based on histopathology images. Although ST data offers valuable insights into the micro-environment of tumors, its acquisition cost remains expensive. Therefore, directly predicting the ST expressions from digital pathology images is desired. Current methods usually adopt existing regression backbones along with patch-sampling for this task, which ignores the inherent multi-scale information embedded in the pyramidal data structure of digital pathology images, and wastes the inter-spot visual information crucial for accurate gene expression prediction. To address these limitations, we propose M2OST, a many-to-one regression Transformer that can accommodate the hierarchical structure of the pathology images via a decoupled multi-scale feature extractor. Unlike traditional models that are trained with one-to-one image-label pairs, M2OST uses multiple images from different levels of the digital pathology image to jointly predict the gene expressions in their common corresponding spot. Built upon our many-to-one scheme, M2OST can be easily scaled to fit different numbers of inputs, and its network structure inherently incorporates nearby inter-spot features, enhancing regression performance. We have tested M2OST on three public ST datasets and the experimental results show that M2OST can achieve state-of-the-art performance with fewer parameters and floating-point operations (FLOPs). The code is available at: https://github.com/Dootmaan/M2OST.
comment: Accepted by AAAI 2025. arXiv admin note: substantial text overlap with arXiv:2401.10608
♻ ☆ Debiased Multimodal Understanding for Human Language Sequences AAAI2025
Multimodal intention understanding (MIU) is an indispensable component of human expression analysis (e.g., sentiment or humor) from heterogeneous modalities, including visual postures, linguistic contents, and acoustic behaviors. Existing works invariably focus on designing sophisticated structures or fusion strategies to achieve impressive improvements. Unfortunately, they all suffer from the subject variation problem due to data distribution discrepancies among subjects. Concretely, MIU models are easily misled by distinct subjects with different expression customs and characteristics in the training data to learn subject-specific spurious correlations, significantly limiting performance and generalizability across uninitiated subjects.Motivated by this observation, we introduce a recapitulative causal graph to formulate the MIU procedure and analyze the confounding effect of subjects. Then, we propose SuCI, a simple yet effective causal intervention module to disentangle the impact of subjects acting as unobserved confounders and achieve model training via true causal effects. As a plug-and-play component, SuCI can be widely applied to most methods that seek unbiased predictions. Comprehensive experiments on several MIU benchmarks clearly demonstrate the effectiveness of the proposed module.
comment: Accepted by AAAI2025
♻ ☆ How secure is AI-generated Code: A Large-Scale Comparison of Large Language Models
This study compares state-of-the-art Large Language Models (LLMs) on their tendency to generate vulnerabilities when writing C programs using a neutral zero-shot prompt. Tihanyi et al. introduced the FormAI dataset at PROMISE'23, featuring 112,000 C programs generated by GPT-3.5-turbo, with over 51.24% identified as vulnerable. We extended that research with a large-scale study involving 9 state-of-the-art models such as OpenAI's GPT-4o-mini, Google's Gemini Pro 1.0, TII's 180 billion-parameter Falcon, Meta's 13 billion-parameter Code Llama, and several other compact models. Additionally, we introduce the FormAI-v2 dataset, which comprises 331 000 compilable C programs generated by these LLMs. Each program in the dataset is labeled based on the vulnerabilities detected in its source code through formal verification, using the Efficient SMT-based Context-Bounded Model Checker (ESBMC). This technique minimizes false positives by providing a counterexample for the specific vulnerability and reduces false negatives by thoroughly completing the verification process. Our study reveals that at least 62.07% of the generated programs are vulnerable. The differences between the models are minor, as they all show similar coding errors with slight variations. Our research highlights that while LLMs offer promising capabilities for code generation, deploying their output in a production environment requires proper risk assessment and validation.
comment: Accepted and will be shortly published at Empirical Software Engineering (EMSE). Journal Impact Factor: 3.5 (2023)
♻ ☆ Semantic Loss Guided Data Efficient Supervised Fine Tuning for Safe Responses in LLMs
Large Language Models (LLMs) generating unsafe responses to toxic prompts is a significant issue in their applications. While various efforts aim to address this safety concern, previous approaches often demand substantial human data collection or rely on the less dependable option of using another LLM to generate corrective data. In this paper, we aim to take this problem and overcome limitations of requiring significant high-quality human data. Our method requires only a small set of unsafe responses to toxic prompts, easily obtained from the unsafe LLM itself. By employing a semantic cost combined with a negative Earth Mover Distance (EMD) loss, we guide the LLM away from generating unsafe responses. Additionally, we propose a novel lower bound for EMD loss, enabling more efficient optimization. Our results demonstrate superior performance and data efficiency compared to baselines, and we further examine the nuanced effects of over-alignment and potential degradation of language capabilities when using contrastive data.
♻ ☆ Uncovering Capabilities of Model Pruning in Graph Contrastive Learning
Graph contrastive learning has achieved great success in pre-training graph neural networks without ground-truth labels. Leading graph contrastive learning follows the classical scheme of contrastive learning, forcing model to identify the essential information from augmented views. However, general augmented views are produced via random corruption or learning, which inevitably leads to semantics alteration. Although domain knowledge guided augmentations alleviate this issue, the generated views are domain specific and undermine the generalization. In this work, motivated by the firm representation ability of sparse model from pruning, we reformulate the problem of graph contrastive learning via contrasting different model versions rather than augmented views. We first theoretically reveal the superiority of model pruning in contrast to data augmentations. In practice, we take original graph as input and dynamically generate a perturbed graph encoder to contrast with the original encoder by pruning its transformation weights. Furthermore, considering the integrity of node embedding in our method, we are capable of developing a local contrastive loss to tackle the hard negative samples that disturb the model training. We extensively validate our method on various benchmarks regarding graph classification via unsupervised and transfer learning. Compared to the state-of-the-art (SOTA) works, better performance can always be obtained by the proposed method.
comment: MM' 24
♻ ☆ HonestLLM: Toward an Honest and Helpful Large Language Model
Large Language Models (LLMs) have achieved remarkable success across various industries due to their exceptional generative capabilities. However, for safe and effective real-world deployments, ensuring honesty and helpfulness is critical. This paper addresses the question: Can we prioritize the helpfulness of LLMs while preserving their honesty? To begin with, we establish exhaustive principles aimed at guaranteeing the honesty of LLM. Additionally, we introduce a novel dataset, referred to as HoneSet, comprising 930 queries spanning six categories meticulously crafted to assess an LLM's capacity for maintaining honesty. Subsequently, we present two approaches to augmenting honesty and helpfulness in LLMs: a training-free enhancement and a fine-tuning-based improvement. The training-free approach, which is based on curiosity-driven prompting, empowers LLMs to articulate internal confusion and uncertainty regarding queries, thereby optimizing their responses. Conversely, the fine-tuning-based method employs a two-stage process inspired by curriculum learning: initially instructing LLMs to discern between honest and dishonest responses, then refining their training to enhance helpfulness. Experiments conducted on nine prominent LLMs demonstrate a significant improvement in alignment with honesty across all models through the implementation of our proposed enhancements. Particularly noteworthy is the 65.3% enhancement observed in Llama3-8b and the remarkable 124.7% improvement in Mistral-7b, as measured by the H$^{2}$ (honest and helpful) assessment. We believe that our work can pave the way for developing more trustworthy LLMs for real-world applications.
♻ ☆ SpecRover: Code Intent Extraction via LLMs ICSE 2025
Autonomous program improvement typically involves automatically producing bug fixes and feature additions. Such program improvement can be accomplished by a combination of large language model (LLM) and program analysis capabilities, in the form of an LLM agent. Since program repair or program improvement typically requires a specification of intended behavior - specification inference can be useful for producing high quality program patches. In this work, we examine efficient and low-cost workflows for iterative specification inference within an LLM agent. Given a GitHub issue to be resolved in a software project, our goal is to conduct iterative code search accompanied by specification inference - thereby inferring intent from both the project structure and behavior. The intent thus captured is examined by a reviewer agent with the goal of vetting the patches as well as providing a measure of confidence in the vetted patches. Our approach SpecRover (AutoCodeRover-v2) is built on the open-source LLM agent AutoCodeRover. In an evaluation on the full SWE-Bench consisting of 2294 GitHub issues, it shows more than 50% improvement in efficacy over AutoCodeRover. Compared to the open-source agents available, our work shows modest cost ($0.65 per issue) in resolving an average GitHub issue in SWE-Bench lite. The production of explanation by SpecRover allows for a better "signal" to be given to the developer, on when the suggested patches can be accepted with confidence. SpecRover also seeks to demonstrate the continued importance of specification inference in automated program repair, even as program repair technologies enter the LLM era.
comment: Haifeng Ruan and Yuntong Zhang contributed equally to this work. To appear in ICSE 2025
♻ ☆ LA4SR: illuminating the dark proteome with generative AI
AI language models (LMs) show promise for biological sequence analysis. We re-engineered open-source LMs (GPT-2, BLOOM, DistilRoBERTa, ELECTRA, and Mamba, ranging from 70M to 12B parameters) for microbial sequence classification. The models achieved F1 scores up to 95 and operated 16,580x faster and at 2.9x the recall of BLASTP. They effectively classified the algal dark proteome - uncharacterized proteins comprising about 65% of total proteins - validated on new data including a new, complete Hi-C/Pacbio Chlamydomonas genome. Larger (>1B) LA4SR models reached high accuracy (F1 > 86) when trained on less than 2% of available data, rapidly achieving strong generalization capacity. High accuracy was achieved when training data had intact or scrambled terminal information, demonstrating robust generalization to incomplete sequences. Finally, we provide custom AI explainability software tools for attributing amino acid patterns to AI generative processes and interpret their outputs in evolutionary and biophysical contexts.
♻ ☆ Improving Mortality Prediction After Radiotherapy with Large Language Model Structuring of Large-Scale Unstructured Electronic Health Records
Accurate survival prediction in radiotherapy (RT) is critical for optimizing treatment decisions. This study developed and validated the RT-Surv framework, which integrates general-domain, open-source large language models (LLMs) to structure unstructured electronic health records alongside structured clinical data. Using data from 34,276 patients and an external cohort of 852, the framework successfully transformed unstructured clinical information into structured formats. Incorporating LLM-structured clinical features improved the concordance index from 0.779 to 0.842 during external validation, demonstrating a significant performance enhancement. Key LLM-structured features, such as disease extent, general condition, and RT purpose, showed high predictive importance and aligned closely with statistically significant predictors identified through conventional statistical analyses, thereby improving model interpretability. Furthermore, the framework enhanced risk stratification, enabling more distinct differentiation among low-, intermediate-, and high-risk groups (p < 0.001) using LLM-structured clinical features. These findings highlight the potential of LLMs to convert unstructured data into actionable insights, improving predictive modeling and patient outcomes in clinics.
comment: 23 pages, 2 tables, 4 figures
♻ ☆ DuoLift-GAN:Reconstructing CT from Single-view and Biplanar X-Rays with Generative Adversarial Networks
Computed tomography (CT) provides highly detailed three-dimensional (3D) medical images but is costly, time-consuming, and often inaccessible in intraoperative settings (Organization et al. 2011). Recent advancements have explored reconstructing 3D chest volumes from sparse 2D X-rays, such as single-view or orthogonal double-view images. However, current models tend to process 2D images in a planar manner, prioritizing visual realism over structural accuracy. In this work, we introduce DuoLift Generative Adversarial Networks (DuoLift-GAN), a novel architecture with dual branches that independently elevate 2D images and their features into 3D representations. These 3D outputs are merged into a unified 3D feature map and decoded into a complete 3D chest volume, enabling richer 3D information capture. We also present a masked loss function that directs reconstruction towards critical anatomical regions, improving structural accuracy and visual quality. This paper demonstrates that DuoLift-GAN significantly enhances reconstruction accuracy while achieving superior visual realism compared to existing methods.
comment: 9 pages, LaTeX; removed the superscript numbers associated with the authors' names for clarity, typos corrected
♻ ☆ Using Game Play to Investigate Multimodal and Conversational Grounding in Large Multimodal Models COLING 2025
While the situation has improved for text-only models, it again seems to be the case currently that multimodal (text and image) models develop faster than ways to evaluate them. In this paper, we bring a recently developed evaluation paradigm from text models to multimodal models, namely evaluation through the goal-oriented game (self) play, complementing reference-based and preference-based evaluation. Specifically, we define games that challenge a model's capability to represent a situation from visual information and align such representations through dialogue. We find that the largest closed models perform rather well on the games that we define, while even the best open-weight models struggle with them. On further analysis, we find that the exceptional deep captioning capabilities of the largest models drive some of the performance. There is still room to grow for both kinds of models, ensuring the continued relevance of the benchmark.
comment: Accepted at COLING 2025
♻ ☆ Improving Deep Reinforcement Learning by Reducing the Chain Effect of Value and Policy Churn NeurIPS 2024
Deep neural networks provide Reinforcement Learning (RL) powerful function approximators to address large-scale decision-making problems. However, these approximators introduce challenges due to the non-stationary nature of RL training. One source of the challenges in RL is that output predictions can churn, leading to uncontrolled changes after each batch update for states not included in the batch. Although such a churn phenomenon exists in each step of network training, how churn occurs and impacts RL remains under-explored. In this work, we start by characterizing churn in a view of Generalized Policy Iteration with function approximation, and we discover a chain effect of churn that leads to a cycle where the churns in value estimation and policy improvement compound and bias the learning dynamics throughout the iteration. Further, we concretize the study and focus on the learning issues caused by the chain effect in different settings, including greedy action deviation in value-based methods, trust region violation in proximal policy optimization, and dual bias of policy value in actor-critic methods. We then propose a method to reduce the chain effect across different settings, called Churn Approximated ReductIoN (CHAIN), which can be easily plugged into most existing DRL algorithms. Our experiments demonstrate the effectiveness of our method in both reducing churn and improving learning performance across online and offline, value-based and policy-based RL settings, as well as a scaling setting.
comment: Accepted to NeurIPS 2024. Project page: https://bluecontra.github.io/CHAIN
♻ ☆ ChatGPT and biometrics: an assessment of face recognition, gender detection, and age estimation capabilities ICIP
This paper explores the application of large language models (LLMs), like ChatGPT, for biometric tasks. We specifically examine the capabilities of ChatGPT in performing biometric-related tasks, with an emphasis on face recognition, gender detection, and age estimation. Since biometrics are considered as sensitive information, ChatGPT avoids answering direct prompts, and thus we crafted a prompting strategy to bypass its safeguard and evaluate the capabilities for biometrics tasks. Our study reveals that ChatGPT recognizes facial identities and differentiates between two facial images with considerable accuracy. Additionally, experimental results demonstrate remarkable performance in gender detection and reasonable accuracy for the age estimation tasks. Our findings shed light on the promising potentials in the application of LLMs and foundation models for biometrics.
comment: Published as a conference paper at IEEE International Conference on Image Processing (ICIP) 2024
♻ ☆ EvolvED: Evolutionary Embeddings to Understand the Generation Process of Diffusion Models
Diffusion models, widely used in image generation, rely on iterative refinement to generate images from noise. Understanding this data evolution is important for model development and interpretability, yet challenging due to its high-dimensional, iterative nature. Prior works often focus on static or instance-level analyses, missing the iterative and holistic aspects of the generative path. While dimensionality reduction can visualize image evolution for few instances, it does preserve the iterative structure. To address these gaps, we introduce EvolvED, a method that presents a holistic view of the iterative generative process in diffusion models. EvolvED goes beyond instance exploration by leveraging predefined research questions to streamline generative space exploration. Tailored prompts aligned with these questions are used to extract intermediate images, preserving iterative context. Targeted feature extractors trace the evolution of key image attribute evolution, addressing the complexity of high-dimensional outputs. Central to EvolvED is a novel evolutionary embedding algorithm that encodes iterative steps while maintaining semantic relations. It enhances the visualization of data evolution by clustering semantically similar elements within each iteration with t-SNE, grouping elements by iteration, and aligning an instance's elements across iterations. We present rectilinear and radial layouts to represent iterations and support exploration. We apply EvolvED to diffusion models like GLIDE and Stable Diffusion, demonstrating its ability to provide valuable insights into the generative process.
♻ ☆ CNNSum: Exploring Long-Context Summarization with Large Language Models in Chinese Novels
Large Language Models (LLMs) have been well-researched in many long-context tasks. However, due to high annotation costs, high-quality long-context summary datasets for training or evaluation are scarce, limiting further research. In this work, we introduce CNNSum, a new multi-scale Chinese long-context novel summarization benchmark, including four subsets, length covering 16k to 128k, 695 samples in total, the annotations are human-driven. We evaluate commercial and open-source models on CNNSum and conduct a detailed analysis. Based on the observations, we further conduct fine-tuning exploration with short-context summary data. In our study: (1) GPT-4o underperformed, due to excessive subjective commentary. (2) Currently, long-context summarization mainly relies on memory ability, small LLMs with stable longer context lengths are the most cost-effective. Using long data concatenated from short-context summaries makes a significant improvement. (3) Prompt templates may cause a large performance gap but can be mitigated through fine-tuning. (4) Fine-tuned Chat or Instruction versions may harm the Base model and further fine-tuning cannot bridge performance gap. (5) while models with RoPE base scaling exhibit strong extrapolation potential, their performance may vary significantly when combined with other interpolation methods and need careful selection. (6) CNNSum provides more reliable and insightful evaluation results than other benchmarks. We release CNNSum to advance research in this field (https://github.com/CxsGhost/CNNSum).
♻ ☆ Mjolnir: Breaking the Shield of Perturbation-Protected Gradients via Adaptive Diffusion AAAI 2025
Perturbation-based mechanisms, such as differential privacy, mitigate gradient leakage attacks by introducing noise into the gradients, thereby preventing attackers from reconstructing clients' private data from the leaked gradients. However, can gradient perturbation protection mechanisms truly defend against all gradient leakage attacks? In this paper, we present the first attempt to break the shield of gradient perturbation protection in Federated Learning for the extraction of private information. We focus on common noise distributions, specifically Gaussian and Laplace, and apply our approach to DNN and CNN models. We introduce Mjolnir, a perturbation-resilient gradient leakage attack that is capable of removing perturbations from gradients without requiring additional access to the original model structure or external data. Specifically, we leverage the inherent diffusion properties of gradient perturbation protection to develop a novel diffusion-based gradient denoising model for Mjolnir. By constructing a surrogate client model that captures the structure of perturbed gradients, we obtain crucial gradient data for training the diffusion model. We further utilize the insight that monitoring disturbance levels during the reverse diffusion process can enhance gradient denoising capabilities, allowing Mjolnir to generate gradients that closely approximate the original, unperturbed versions through adaptive sampling steps. Extensive experiments demonstrate that Mjolnir effectively recovers the protected gradients and exposes the Federated Learning process to the threat of gradient leakage, achieving superior performance in gradient denoising and private data recovery.
comment: Accepted by AAAI 2025
♻ ☆ Language Ranker: A Metric for Quantifying LLM Performance Across High and Low-Resource Languages AAAI 2025
The development of Large Language Models (LLMs) relies on extensive text corpora, which are often unevenly distributed across languages. This imbalance results in LLMs performing significantly better on high-resource languages like English, German, and French, while their capabilities in low-resource languages remain inadequate. Currently, there is a lack of quantitative methods to evaluate the performance of LLMs in these low-resource languages. To address this gap, we propose the Language Ranker, an intrinsic metric designed to benchmark and rank languages based on LLM performance using internal representations. By comparing the LLM's internal representation of various languages against a baseline derived from English, we can assess the model's multilingual capabilities in a robust and language-agnostic manner. Our analysis reveals that high-resource languages exhibit higher similarity scores with English, demonstrating superior performance, while low-resource languages show lower similarity scores, underscoring the effectiveness of our metric in assessing language-specific capabilities. Besides, the experiments show that there is a strong correlation between the LLM's performance in different languages and the proportion of those languages in its pre-training corpus. These insights underscore the efficacy of the Language Ranker as a tool for evaluating LLM performance across different languages, particularly those with limited resources.
comment: Accepted by AAAI 2025 (Social Impact Track)
♻ ☆ Hidden in the Noise: Two-Stage Robust Watermarking for Images
As the quality of image generators continues to improve, deepfakes become a topic of considerable societal debate. Image watermarking allows responsible model owners to detect and label their AI-generated content, which can mitigate the harm. Yet, current state-of-the-art methods in image watermarking remain vulnerable to forgery and removal attacks. This vulnerability occurs in part because watermarks distort the distribution of generated images, unintentionally revealing information about the watermarking techniques. In this work, we first demonstrate a distortion-free watermarking method for images, based on a diffusion model's initial noise. However, detecting the watermark requires comparing the initial noise reconstructed for an image to all previously used initial noises. To mitigate these issues, we propose a two-stage watermarking framework for efficient detection. During generation, we augment the initial noise with generated Fourier patterns to embed information about the group of initial noises we used. For detection, we (i) retrieve the relevant group of noises, and (ii) search within the given group for an initial noise that might match our image. This watermarking approach achieves state-of-the-art robustness to forgery and removal against a large battery of attacks.
♻ ☆ Representation Learning with Large Language Models for Recommendation WWW'24
Recommender systems have seen significant advancements with the influence of deep learning and graph neural networks, particularly in capturing complex user-item relationships. However, these graph-based recommenders heavily depend on ID-based data, potentially disregarding valuable textual information associated with users and items, resulting in less informative learned representations. Moreover, the utilization of implicit feedback data introduces potential noise and bias, posing challenges for the effectiveness of user preference learning. While the integration of large language models (LLMs) into traditional ID-based recommenders has gained attention, challenges such as scalability issues, limitations in text-only reliance, and prompt input constraints need to be addressed for effective implementation in practical recommender systems. To address these challenges, we propose a model-agnostic framework RLMRec that aims to enhance existing recommenders with LLM-empowered representation learning. It proposes a recommendation paradigm that integrates representation learning with LLMs to capture intricate semantic aspects of user behaviors and preferences. RLMRec incorporates auxiliary textual signals, develops a user/item profiling paradigm empowered by LLMs, and aligns the semantic space of LLMs with the representation space of collaborative relational signals through a cross-view alignment framework. This work further establish a theoretical foundation demonstrating that incorporating textual signals through mutual information maximization enhances the quality of representations. In our evaluation, we integrate RLMRec with state-of-the-art recommender models, while also analyzing its efficiency and robustness to noise data. Our implementation codes are available at https://github.com/HKUDS/RLMRec.
comment: Published as a WWW'24 full paper
♻ ☆ Towards Data-centric Machine Learning on Directed Graphs: a Survey
In recent years, Graph Neural Networks (GNNs) have made significant advances in processing structured data. However, most of them primarily adopted a model-centric approach, which simplifies graphs by converting them into undirected formats and emphasizes model designs. This approach is inherently limited in real-world applications due to the unavoidable information loss in simple undirected graphs and the model optimization challenges that arise when exceeding the upper bounds of this sub-optimal data representational capacity. As a result, there has been a shift toward data-centric methods that prioritize improving graph quality and representation. Specifically, various types of graphs can be derived from naturally structured data, including heterogeneous graphs, hypergraphs, and directed graphs. Among these, directed graphs offer distinct advantages in topological systems by modeling causal relationships, and directed GNNs have been extensively studied in recent years. However, a comprehensive survey of this emerging topic is still lacking. Therefore, we aim to provide a comprehensive review of directed graph learning, with a particular focus on a data-centric perspective. Specifically, we first introduce a novel taxonomy for existing studies. Subsequently, we re-examine these methods from the data-centric perspective, with an emphasis on understanding and improving data representation. It demonstrates that a deep understanding of directed graphs and their quality plays a crucial role in model performance. Additionally, we explore the diverse applications of directed GNNs across 10+ domains, highlighting their broad applicability. Finally, we identify key opportunities and challenges within the field, offering insights that can guide future research and development in directed graph learning.
comment: In Progress
♻ ☆ Efficient Prompt Tuning by Multi-Space Projection and Prompt Fusion
Prompt tuning is a promising method to fine-tune a pre-trained language model without retraining its large-scale parameters. Instead, it attaches a soft prompt to the input text, whereby downstream tasks can be well adapted by merely learning the embeddings of prompt tokens. Nevertheless, existing methods still suffer from two challenges: (i) they are hard to balance accuracy and efficiency. A longer (shorter) soft prompt generally leads to a better(worse) accuracy but at the cost of more (less) training time. (ii)The performance may not be consistent when adapting to different downstream tasks. We attribute it to the same embedding space but responsible for different requirements of downstream tasks. To address these issues, we propose an Efficient Prompt Tuning method (EPT) by multi-space projection and prompt fusion. Specifically, it decomposes a given soft prompt into a shorter prompt and two low-rank matrices, significantly reducing the training time. Accuracy is also enhanced by leveraging low-rank matrices and the short prompt as additional knowledge sources to enrich the semantics of the original short prompt. In addition, we project the soft prompt into multiple subspaces to improve the performance consistency, and then adaptively learn the combination weights of different spaces through a gating network. Experiments on 13 natural language processing downstream tasks show that our method significantly and consistently outperforms 11 comparison methods with the relative percentage of improvements up to 12.9%, and training time decreased by 14%.
♻ ☆ Scaling Up Probabilistic Circuits by Latent Variable Distillation
Probabilistic Circuits (PCs) are a unified framework for tractable probabilistic models that support efficient computation of various probabilistic queries (e.g., marginal probabilities). One key challenge is to scale PCs to model large and high-dimensional real-world datasets: we observe that as the number of parameters in PCs increases, their performance immediately plateaus. This phenomenon suggests that the existing optimizers fail to exploit the full expressive power of large PCs. We propose to overcome such bottleneck by latent variable distillation: we leverage the less tractable but more expressive deep generative models to provide extra supervision over the latent variables of PCs. Specifically, we extract information from Transformer-based generative models to assign values to latent variables of PCs, providing guidance to PC optimizers. Experiments on both image and language modeling benchmarks (e.g., ImageNet and WikiText-2) show that latent variable distillation substantially boosts the performance of large PCs compared to their counterparts without latent variable distillation. In particular, on the image modeling benchmarks, PCs achieve competitive performance against some of the widely-used deep generative models, including variational autoencoders and flow-based models, opening up new avenues for tractable generative modeling. Our code can be found at https://github.com/UCLA-StarAI/LVD.
♻ ☆ Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space NeurIPS 2024
Modern generative models demonstrate impressive capabilities, likely stemming from an ability to identify and manipulate abstract concepts underlying their training data. However, fundamental questions remain: what determines the concepts a model learns, the order in which it learns them, and its ability to manipulate those concepts? To address these questions, we propose analyzing a model's learning dynamics via a framework we call the concept space, where each axis represents an independent concept underlying the data generating process. By characterizing learning dynamics in this space, we identify how the speed at which a concept is learned, and hence the order of concept learning, is controlled by properties of the data we term concept signal. Further, we observe moments of sudden turns in the direction of a model's learning dynamics in concept space. Surprisingly, these points precisely correspond to the emergence of hidden capabilities, i.e., where latent interventions show the model possesses the capability to manipulate a concept, but these capabilities cannot yet be elicited via naive input prompting. While our results focus on synthetically defined toy datasets, we hypothesize a general claim on emergence of hidden capabilities may hold: generative models possess latent capabilities that emerge suddenly and consistently during training, though a model might not exhibit these capabilities under naive input prompting.
comment: NeurIPS 2024 (Spotlight)
♻ ☆ GameArena: Evaluating LLM Reasoning through Live Computer Games
Evaluating the reasoning abilities of large language models (LLMs) is challenging. Existing benchmarks often depend on static datasets, which are vulnerable to data contamination and may get saturated over time, or on binary live human feedback that conflates reasoning with other abilities. As the most prominent dynamic benchmark, Chatbot Arena evaluates open-ended questions in real-world settings, but lacks the granularity in assessing specific reasoning capabilities. We introduce GameArena, a dynamic benchmark designed to evaluate LLM reasoning capabilities through interactive gameplay with humans. GameArena consists of three games designed to test specific reasoning capabilities (e.g., deductive and inductive reasoning), while keeping participants entertained and engaged. We analyze the gaming data retrospectively to uncover the underlying reasoning processes of LLMs and measure their fine-grained reasoning capabilities. We collect over 2000 game sessions and provide detailed assessments of various reasoning capabilities for five state-of-the-art LLMs. Our user study with 100 participants suggests that GameArena improves user engagement compared to Chatbot Arena. For the first time, GameArena enables the collection of step-by-step LLM reasoning data in the wild.
♻ ☆ CohEx: A Generalized Framework for Cohort Explanation
eXplainable Artificial Intelligence (XAI) has garnered significant attention for enhancing transparency and trust in machine learning models. However, the scopes of most existing explanation techniques focus either on offering a holistic view of the explainee model (global explanation) or on individual instances (local explanation), while the middle ground, i.e., cohort-based explanation, is less explored. Cohort explanations offer insights into the explainee's behavior on a specific group or cohort of instances, enabling a deeper understanding of model decisions within a defined context. In this paper, we discuss the unique challenges and opportunities associated with measuring cohort explanations, define their desired properties, and create a generalized framework for generating cohort explanations based on supervised clustering.
♻ ☆ Tube Loss: A Novel Approach for Prediction Interval Estimation and probabilistic forecasting
This paper proposes a novel loss function, called 'Tube Loss', for simultaneous estimation of bounds of a Prediction Interval (PI) in the regression setup, and also for generating probabilistic forecasts from time series data solving a single optimization problem. The PIs obtained by minimizing the empirical risk based on the Tube Loss are shown to be of better quality than the PIs obtained by the existing methods in the following sense. First, it yields intervals that attain the prespecified confidence level $t \in(0,1)$ asymptotically. A theoretical proof of this fact is given. Secondly, the user is allowed to move the interval up or down by controlling the value of a parameter. This helps the user to choose a PI capturing denser regions of the probability distribution of the response variable inside the interval, and thus, sharpening its width. This is shown to be especially useful when the conditional distribution of the response variable is skewed. Further, the Tube Loss based PI estimation method can trade-off between the coverage and the average width by solving a single optimization problem. It enables further reduction of the average width of PI through re-calibration. Also, unlike a few existing PI estimation methods the gradient descent (GD) method can be used for minimization of empirical risk. Finally, through extensive experimentation, we have shown the efficacy of the Tube Loss based PI estimation in kernel machines, neural networks and deep networks and also for probabilistic forecasting tasks. The codes of the experiments are available at https://github.com/ltpritamanand/Tube_loss
♻ ☆ TorchSpatial: A Location Encoding Framework and Benchmark for Spatial Representation Learning NeurIPS 2024
Spatial representation learning (SRL) aims at learning general-purpose neural network representations from various types of spatial data (e.g., points, polylines, polygons, networks, images, etc.) in their native formats. Learning good spatial representations is a fundamental problem for various downstream applications such as species distribution modeling, weather forecasting, trajectory generation, geographic question answering, etc. Even though SRL has become the foundation of almost all geospatial artificial intelligence (GeoAI) research, we have not yet seen significant efforts to develop an extensive deep learning framework and benchmark to support SRL model development and evaluation. To fill this gap, we propose TorchSpatial, a learning framework and benchmark for location (point) encoding, which is one of the most fundamental data types of spatial representation learning. TorchSpatial contains three key components: 1) a unified location encoding framework that consolidates 15 commonly recognized location encoders, ensuring scalability and reproducibility of the implementations; 2) the LocBench benchmark tasks encompassing 7 geo-aware image classification and 10 geo-aware image regression datasets; 3) a comprehensive suite of evaluation metrics to quantify geo-aware models' overall performance as well as their geographic bias, with a novel Geo-Bias Score metric. Finally, we provide a detailed analysis and insights into the model performance and geographic bias of different location encoders. We believe TorchSpatial will foster future advancement of spatial representation learning and spatial fairness in GeoAI research. The TorchSpatial model framework, LocBench, and Geo-Bias Score evaluation framework are available at https://github.com/seai-lab/TorchSpatial.
comment: 10 pages, 2 figures. Accepted by NeurIPS 2024 Datasets and Benchmarks Track
♻ ☆ Reducing Inference Energy Consumption Using Dual Complementary CNNs
Energy efficiency of Convolutional Neural Networks (CNNs) has become an important area of research, with various strategies being developed to minimize the power consumption of these models. Previous efforts, including techniques like model pruning, quantization, and hardware optimization, have made significant strides in this direction. However, there remains a need for more effective on device AI solutions that balance energy efficiency with model performance. In this paper, we propose a novel approach to reduce the energy requirements of inference of CNNs. Our methodology employs two small Complementary CNNs that collaborate with each other by covering each other's "weaknesses" in predictions. If the confidence for a prediction of the first CNN is considered low, the second CNN is invoked with the aim of producing a higher confidence prediction. This dual-CNN setup significantly reduces energy consumption compared to using a single large deep CNN. Additionally, we propose a memory component that retains previous classifications for identical inputs, bypassing the need to re-invoke the CNNs for the same input, further saving energy. Our experiments on a Jetson Nano computer demonstrate an energy reduction of up to 85.8% achieved on modified datasets where each sample was duplicated once. These findings indicate that leveraging a complementary CNN pair along with a memory component effectively reduces inference energy while maintaining high accuracy.
♻ ☆ Exploring Consistency in Graph Representations:from Graph Kernels to Graph Neural Networks NeurIPS 2024
Graph Neural Networks (GNNs) have emerged as a dominant approach in graph representation learning, yet they often struggle to capture consistent similarity relationships among graphs. While graph kernel methods such as the Weisfeiler-Lehman subtree (WL-subtree) and Weisfeiler-Lehman optimal assignment (WLOA) kernels are effective in capturing similarity relationships, they rely heavily on predefined kernels and lack sufficient non-linearity for more complex data patterns. Our work aims to bridge the gap between neural network methods and kernel approaches by enabling GNNs to consistently capture relational structures in their learned representations. Given the analogy between the message-passing process of GNNs and WL algorithms, we thoroughly compare and analyze the properties of WL-subtree and WLOA kernels. We find that the similarities captured by WLOA at different iterations are asymptotically consistent, ensuring that similar graphs remain similar in subsequent iterations, thereby leading to superior performance over the WL-subtree kernel. Inspired by these findings, we conjecture that the consistency in the similarities of graph representations across GNN layers is crucial in capturing relational structures and enhancing graph classification performance. Thus, we propose a loss to enforce the similarity of graph representations to be consistent across different layers. Our empirical analysis verifies our conjecture and shows that our proposed consistency loss can significantly enhance graph classification performance across several GNN backbones on various datasets.
comment: NeurIPS 2024
♻ ☆ Dynamic Universal Approximation Theory: The Basic Theory for Transformer-based Large Language Models
Language models have emerged as a critical area of focus in artificial intelligence, particularly with the introduction of groundbreaking innovations like ChatGPT. Large-scale Transformer networks have quickly become the leading approach for advancing natural language processing algorithms. Built on the Transformer architecture, these models enable interactions that closely mimic human communication and, equipped with extensive knowledge, can even assist in guiding human tasks. Despite their impressive capabilities and growing complexity, a key question remains-the theoretical foundations of large language models (LLMs). What makes Transformer so effective for powering intelligent language applications, such as translation and coding? What underlies LLMs' ability for In-Context Learning (ICL)? How does the LoRA scheme enhance the fine-tuning of LLMs? And what supports the practicality of pruning LLMs? To address these critical questions and explore the technological strategies within LLMs, we leverage the Universal Approximation Theory (UAT) to offer a theoretical backdrop, shedding light on the mechanisms that underpin these advancements.
♻ ☆ Dynamic Demonstrations Controller for In-Context Learning
In-context learning (ICL) is a new paradigm for natural language processing (NLP), where a large language model (LLM) observes a small number of demonstrations and a test instance as its input, and directly makes predictions without updating model parameters. Previous studies have revealed that ICL is sensitive to the selection and the ordering of demonstrations. However, there are few studies regarding the impact of the demonstration number on the ICL performance within a limited input length of LLM, because it is commonly believed that the number of demonstrations is positively correlated with model performance. In this paper, we found this conclusion does not always hold true. Through pilot experiments, we discover that increasing the number of demonstrations does not necessarily lead to improved performance. Building upon this insight, we propose a Dynamic Demonstrations Controller (D$^2$Controller), which can improve the ICL performance by adjusting the number of demonstrations dynamically. The experimental results show that D$^2$Controller yields a 4.6% relative improvement on ten different sizes of LLMs across ten datasets. Moreover, we also extend our method to previous ICL models and achieve competitive results.
♻ ☆ Conditioned quantum-assisted deep generative surrogate for particle-calorimeter interactions
Particle collisions at accelerators such as the Large Hadron Collider, recorded and analyzed by experiments such as ATLAS and CMS, enable exquisite measurements of the Standard Model and searches for new phenomena. Simulations of collision events at these detectors have played a pivotal role in shaping the design of future experiments and analyzing ongoing ones. However, the quest for accuracy in Large Hadron Collider (LHC) collisions comes at an imposing computational cost, with projections estimating the need for millions of CPU-years annually during the High Luminosity LHC (HL-LHC) run \cite{collaboration2022atlas}. Simulating a single LHC event with \textsc{Geant4} currently devours around 1000 CPU seconds, with simulations of the calorimeter subdetectors in particular imposing substantial computational demands \cite{rousseau2023experimental}. To address this challenge, we propose a conditioned quantum-assisted deep generative model. Our model integrates a conditioned variational autoencoder (VAE) on the exterior with a conditioned Restricted Boltzmann Machine (RBM) in the latent space, providing enhanced expressiveness compared to conventional VAEs. The RBM nodes and connections are meticulously engineered to enable the use of qubits and couplers on D-Wave's Pegasus-structured \textit{Advantage} quantum annealer (QA) for sampling. We introduce a novel method for conditioning the quantum-assisted RBM using \textit{flux biases}. We further propose a novel adaptive mapping to estimate the effective inverse temperature in quantum annealers. The effectiveness of our framework is illustrated using Dataset 2 of the CaloChallenge \cite{calochallenge}.
comment: 27 pages, 10 figures, 8 appendices
♻ ☆ Deep Learning and Machine Learning, Advancing Big Data Analytics and Management: Tensorflow Pretrained Models
The application of TensorFlow pre-trained models in deep learning is explored, with an emphasis on practical guidance for tasks such as image classification and object detection. The study covers modern architectures, including ResNet, MobileNet, and EfficientNet, and demonstrates the effectiveness of transfer learning through real-world examples and experiments. A comparison of linear probing and model fine-tuning is presented, supplemented by visualizations using techniques like PCA, t-SNE, and UMAP, allowing for an intuitive understanding of the impact of these approaches. The work provides complete example code and step-by-step instructions, offering valuable insights for both beginners and advanced users. By integrating theoretical concepts with hands-on practice, the paper equips readers with the tools necessary to address deep learning challenges efficiently.
comment: This book contains 148 pages and 7 figures
♻ ☆ Human-Object Interaction from Human-Level Instructions
Intelligent agents must autonomously interact with the environments to perform daily tasks based on human-level instructions. They need a foundational understanding of the world to accurately interpret these instructions, along with precise low-level movement and interaction skills to execute the derived actions. In this work, we propose the first complete system for synthesizing physically plausible, long-horizon human-object interactions for object manipulation in contextual environments, driven by human-level instructions. We leverage large language models (LLMs) to interpret the input instructions into detailed execution plans. Unlike prior work, our system is capable of generating detailed finger-object interactions, in seamless coordination with full-body movements. We also train a policy to track generated motions in physics simulation via reinforcement learning (RL) to ensure physical plausibility of the motion. Our experiments demonstrate the effectiveness of our system in synthesizing realistic interactions with diverse objects in complex environments, highlighting its potential for real-world applications.
comment: project page: https://hoifhli.github.io/
♻ ☆ Large Language Models and Cognitive Science: A Comprehensive Review of Similarities, Differences, and Challenges
This comprehensive review explores the intersection of Large Language Models (LLMs) and cognitive science, examining similarities and differences between LLMs and human cognitive processes. We analyze methods for evaluating LLMs cognitive abilities and discuss their potential as cognitive models. The review covers applications of LLMs in various cognitive fields, highlighting insights gained for cognitive science research. We assess cognitive biases and limitations of LLMs, along with proposed methods for improving their performance. The integration of LLMs with cognitive architectures is examined, revealing promising avenues for enhancing artificial intelligence (AI) capabilities. Key challenges and future research directions are identified, emphasizing the need for continued refinement of LLMs to better align with human cognition. This review provides a balanced perspective on the current state and future potential of LLMs in advancing our understanding of both artificial and human intelligence.
comment: 10 pages, 1 figure
♻ ☆ Strong but simple: A Baseline for Domain Generalized Dense Perception by CLIP-based Transfer Learning ACCV 2024
Domain generalization (DG) remains a significant challenge for perception based on deep neural networks (DNNs), where domain shifts occur due to synthetic data, lighting, weather, or location changes. Vision-language models (VLMs) marked a large step for the generalization capabilities and have been already applied to various tasks. Very recently, first approaches utilized VLMs for domain generalized segmentation and object detection and obtained strong generalization. However, all these approaches rely on complex modules, feature augmentation frameworks or additional models. Surprisingly and in contrast to that, we found that simple fine-tuning of vision-language pre-trained models yields competitive or even stronger generalization results while being extremely simple to apply. Moreover, we found that vision-language pre-training consistently provides better generalization than the previous standard of vision-only pre-training. This challenges the standard of using ImageNet-based transfer learning for domain generalization. Fully fine-tuning a vision-language pre-trained model is capable of reaching the domain generalization SOTA when training on the synthetic GTA5 dataset. Moreover, we confirm this observation for object detection on a novel synthetic-to-real benchmark. We further obtain superior generalization capabilities by reaching 77.9% mIoU on the popular Cityscapes-to-ACDC benchmark. We also found improved in-domain generalization, leading to an improved SOTA of 86.4% mIoU on the Cityscapes test set marking the first place on the leaderboard.
comment: Accepted to ACCV 2024; Project Page: https://vltseg.github.io/
♻ ☆ Woodpecker: Hallucination Correction for Multimodal Large Language Models SC
Hallucination is a big shadow hanging over the rapidly evolving Multimodal Large Language Models (MLLMs), referring to the phenomenon that the generated text is inconsistent with the image content. In order to mitigate hallucinations, existing studies mainly resort to an instruction-tuning manner that requires retraining the models with specific data. In this paper, we pave a different way, introducing a training-free method named Woodpecker. Like a woodpecker heals trees, it picks out and corrects hallucinations from the generated text. Concretely, Woodpecker consists of five stages: key concept extraction, question formulation, visual knowledge validation, visual claim generation, and hallucination correction. Implemented in a post-remedy manner, Woodpecker can easily serve different MLLMs, while being interpretable by accessing intermediate outputs of the five stages. We evaluate Woodpecker both quantitatively and qualitatively and show the huge potential of this new paradigm. On the POPE benchmark, our method obtains a 30.66%/24.33% improvement in accuracy over the baseline MiniGPT-4/mPLUG-Owl. The source code is released at https://github.com/BradyFU/Woodpecker.
comment: Accepted by Science China Information Sciences (SCIS)
♻ ☆ Enhancing Relation Extraction via Supervised Rationale Verification and Feedback AAAI 2025
Despite the rapid progress that existing automated feedback methods have made in correcting the output of large language models (LLMs), these methods cannot be well applied to the relation extraction (RE) task due to their designated feedback objectives and correction manner. To address this problem, we propose a novel automated feedback framework for RE, which presents a rationale supervisor to verify the rationale and provides re-selected demonstrations as feedback to correct the initial prediction. Specifically, we first design a causal intervention and observation method to collect biased/unbiased rationales for contrastive training the rationale supervisor. Then, we present a verification-feedback-correction procedure to iteratively enhance LLMs' capability of handling the RE task. Extensive experiments prove that our proposed framework significantly outperforms existing methods.
comment: Accepted to AAAI 2025, camera ready version
♻ ☆ Stag-1: Towards Realistic 4D Driving Simulation with Video Generation Model
4D driving simulation is essential for developing realistic autonomous driving simulators. Despite advancements in existing methods for generating driving scenes, significant challenges remain in view transformation and spatial-temporal dynamic modeling. To address these limitations, we propose a Spatial-Temporal simulAtion for drivinG (Stag-1) model to reconstruct real-world scenes and design a controllable generative network to achieve 4D simulation. Stag-1 constructs continuous 4D point cloud scenes using surround-view data from autonomous vehicles. It decouples spatial-temporal relationships and produces coherent keyframe videos. Additionally, Stag-1 leverages video generation models to obtain photo-realistic and controllable 4D driving simulation videos from any perspective. To expand the range of view generation, we train vehicle motion videos based on decomposed camera poses, enhancing modeling capabilities for distant scenes. Furthermore, we reconstruct vehicle camera trajectories to integrate 3D points across consecutive views, enabling comprehensive scene understanding along the temporal dimension. Following extensive multi-level scene training, Stag-1 can simulate from any desired viewpoint and achieve a deep understanding of scene evolution under static spatial-temporal conditions. Compared to existing methods, our approach shows promising performance in multi-view scene consistency, background coherence, and accuracy, and contributes to the ongoing advancements in realistic autonomous driving simulation. Code: https://github.com/wzzheng/Stag.
comment: Code is available at: https://github.com/wzzheng/Stag
♻ ☆ Technical Report: Enhancing LLM Reasoning with Reward-guided Tree Search
Recently, test-time scaling has garnered significant attention from the research community, largely due to the substantial advancements of the o1 model released by OpenAI. By allocating more computational resources during the inference phase, large language models~(LLMs) can extensively explore the solution space by generating more thought tokens or diverse solutions, thereby producing more accurate responses. However, developing an o1-like reasoning approach is challenging, and researchers have been making various attempts to advance this open area of research. In this paper, we present a preliminary exploration into enhancing the reasoning abilities of LLMs through reward-guided tree search algorithms. This framework is implemented by integrating the policy model, reward model, and search algorithm. It is primarily constructed around a tree search algorithm, where the policy model navigates a dynamically expanding tree guided by a specially trained reward model. We thoroughly explore various design considerations necessary for implementing this framework and provide a detailed report of the technical aspects. To assess the effectiveness of our approach, we focus on mathematical reasoning tasks and conduct extensive evaluations on four challenging datasets, significantly enhancing the reasoning abilities of LLMs.
comment: Technical Report on Slow Thinking with LLMs: I
♻ ☆ The EarlyBird Gets the WORM: Heuristically Accelerating EarlyBird Convergence NeurIPS 2024
The Lottery Ticket hypothesis proposes that ideal, sparse subnetworks, called lottery tickets, exist in untrained dense neural networks. The Early Bird hypothesis proposes an efficient algorithm to find these winning lottery tickets in convolutional neural networks, using the novel concept of distance between subnetworks to detect convergence in the subnetworks of a model. However, this approach overlooks unchanging groups of unimportant neurons near the search's end. We proposes WORM, a method that exploits these static groups by truncating their gradients, forcing the model to rely on other neurons. Experiments show WORM achieves faster ticket identification during training on convolutional neural networks, despite the additional computational overhead, when compared to EarlyBird search. Additionally, WORM-pruned models lose less accuracy during pruning and recover accuracy faster, improving the robustness of a given model. Furthermore, WORM is also able to generalize the Early Bird hypothesis reasonably well to larger models, such as transformers, displaying its flexibility to adapt to more complex architectures.
comment: Accepted to the Efficient Natural Language and Speech Processing Workshop at NeurIPS 2024
♻ ☆ Map Optical Properties to Subwavelength Structures Directly via a Diffusion Model
Subwavelength photonic structures and metamaterials provide revolutionary approaches for controlling light. The inverse design methods proposed for these subwavelength structures are vital to the development of new photonic devices. However, most of the existing inverse design methods cannot realize direct mapping from optical properties to photonic structures but instead rely on forward simulation methods to perform iterative optimization. In this work, we exploit the powerful generative abilities of artificial intelligence (AI) and propose a practical inverse design method based on latent diffusion models. Our method maps directly the optical properties to structures without the requirement of forward simulation and iterative optimization. Here, the given optical properties can work as "prompts" and guide the constructed model to correctly "draw" the required photonic structures. Experiments show that our direct mapping-based inverse design method can generate subwavelength photonic structures at high fidelity while following the given optical properties. This may change the method used for optical design and greatly accelerate the research on new photonic devices.
♻ ☆ Structured Unrestricted-Rank Matrices for Parameter Efficient Fine-tuning NeurIPS 2024
Recent efforts to scale Transformer models have demonstrated rapid progress across a wide range of tasks (Wei et al., 2022). However, fine-tuning these models for downstream tasks is expensive due to their large parameter counts. Parameter-efficient fine-tuning (PEFT) approaches have emerged as a viable alternative by allowing us to fine-tune models by updating only a small number of parameters. In this work, we propose a general framework for parameter efficient fine-tuning (PEFT), based on structured unrestricted-rank matrices (SURM) which can serve as a drop-in replacement for popular approaches such as Adapters and LoRA. Unlike other methods like LoRA, SURMs provides more flexibility in finding the right balance between compactness and expressiveness. This is achieved by using low displacement rank matrices (LDRMs), which hasn't been used in this context before. SURMs remain competitive with baselines, often providing significant quality improvements while using a smaller parameter budget. SURMs achieve 5-7% accuracy gains on various image classification tasks while replacing low-rank matrices in LoRA. It also results in up to 12x reduction of the number of parameters in adapters (with virtually no loss in quality) on the GLUE benchmark.
comment: Accepted at NeurIPS 2024. Updated draft at: https://openreview.net/pdf?id=MXOzgjlWDF
♻ ☆ MicroScopiQ: Accelerating Foundational Models through Outlier-Aware Microscaling Quantization
Quantization of foundational models (FMs) is significantly more challenging than traditional DNNs due to the emergence of large magnitude features called outliers. Existing outlier-aware algorithm/architecture co-design techniques either use mixed-precision, retaining outliers at high precision but compromise hardware efficiency, or quantize inliers and outliers at the same precision, improving hardware efficiency at the cost of accuracy. To address this mutual exclusivity, in this paper, we propose MicroScopiQ, a novel co-design technique that leverages pruning to complement outlier-aware quantization. MicroScopiQ retains outliers at higher precision while pruning a certain fraction of least important weights to distribute the additional outlier bits; ensuring high accuracy, aligned memory and hardware efficiency. We design a high-throughput, low overhead accelerator architecture composed of simple multi-precision INT processing elements and a novel network-on-chip called ReCoN that efficiently abstracts the complexity of supporting high-precision outliers. Additionally, unlike existing alternatives, MicroScopiQ does not assume any locality of outlier weights, enabling applicability to a broad range of FMs. Extensive experiments across various quantization settings show that MicroScopiQ achieves SoTA quantization performance while simultaneously improving inference performance by 3x and reducing energy by 2x over existing alternatives.
comment: Preprint under review
Computation and Language 107
☆ Zero-Shot ATC Coding with Large Language Models for Clinical Assessments
Manual assignment of Anatomical Therapeutic Chemical (ATC) codes to prescription records is a significant bottleneck in healthcare research and operations at Ontario Health and InterRAI Canada, requiring extensive expert time and effort. To automate this process while maintaining data privacy, we develop a practical approach using locally deployable large language models (LLMs). Inspired by recent advances in automatic International Classification of Diseases (ICD) coding, our method frames ATC coding as a hierarchical information extraction task, guiding LLMs through the ATC ontology level by level. We evaluate our approach using GPT-4o as an accuracy ceiling and focus development on open-source Llama models suitable for privacy-sensitive deployment. Testing across Health Canada drug product data, the RABBITS benchmark, and real clinical notes from Ontario Health, our method achieves 78% exact match accuracy with GPT-4o and 60% with Llama 3.1 70B. We investigate knowledge grounding through drug definitions, finding modest improvements in accuracy. Further, we show that fine-tuned Llama 3.1 8B matches zero-shot Llama 3.1 70B accuracy, suggesting that effective ATC coding is feasible with smaller models. Our results demonstrate the feasibility of automatic ATC coding in privacy-sensitive healthcare environments, providing a foundation for future deployments.
☆ Granite Guardian
We introduce the Granite Guardian models, a suite of safeguards designed to provide risk detection for prompts and responses, enabling safe and responsible use in combination with any large language model (LLM). These models offer comprehensive coverage across multiple risk dimensions, including social bias, profanity, violence, sexual content, unethical behavior, jailbreaking, and hallucination-related risks such as context relevance, groundedness, and answer relevance for retrieval-augmented generation (RAG). Trained on a unique dataset combining human annotations from diverse sources and synthetic data, Granite Guardian models address risks typically overlooked by traditional risk detection models, such as jailbreaks and RAG-specific issues. With AUC scores of 0.871 and 0.854 on harmful content and RAG-hallucination-related benchmarks respectively, Granite Guardian is the most generalizable and competitive model available in the space. Released as open-source, Granite Guardian aims to promote responsible AI development across the community. https://github.com/ibm-granite/granite-guardian
☆ TRIM: Token Reduction and Inference Modeling for Cost-Effective Language Generation
The inference cost of Large Language Models (LLMs) is a significant challenge due to their computational demands, specially on tasks requiring long outputs. However, natural language often contains redundancy, which presents an opportunity for optimization. We have observed that LLMs can generate distilled language-concise outputs that retain essential meaning, when prompted appropriately. We propose a framework for saving computational cost, in which a shorter distilled output from the LLM is reconstructed into a full narrative by a smaller model with lower inference costs. Our experiments show promising results, particularly in general knowledge domains with 20.58% saved tokens on average with tiny decrease in evaluation metrics, hinting that this approach can effectively balance efficiency and accuracy in language processing tasks.
comment: 12 pages
☆ Can linguists better understand DNA?
Multilingual transfer ability, which reflects how well models fine-tuned on one source language can be applied to other languages, has been well studied in multilingual pre-trained models. However, the existence of such capability transfer between natural language and gene sequences/languages remains underexplored.This study addresses this gap by drawing inspiration from the sentence-pair classification task used for evaluating sentence similarity in natural language. We constructed two analogous tasks: DNA-pair classification(DNA sequence similarity) and DNA-protein-pair classification(gene coding determination). These tasks were designed to validate the transferability of capabilities from natural language to gene sequences. Even a small-scale pre-trained model like GPT-2-small, which was pre-trained on English, achieved an accuracy of 78% on the DNA-pair classification task after being fine-tuned on English sentence-pair classification data(XTREME PAWS-X). While training a BERT model on multilingual text, the precision reached 82%.On the more complex DNA-protein-pair classification task, however, the model's output was barely distinguishable from random output.Experiments suggest that there may be a capability transfer from natural language to genetic language, but further task testing is needed to confirm this.
comment: 11 pages,3 figures
☆ RAZOR: Sharpening Knowledge by Cutting Bias with Unsupervised Text Rewriting AAAI'25
Despite the widespread use of LLMs due to their superior performance in various tasks, their high computational costs often lead potential users to opt for the pretraining-finetuning pipeline. However, biases prevalent in manually constructed datasets can introduce spurious correlations between tokens and labels, creating so-called shortcuts and hindering the generalizability of fine-tuned models. Existing debiasing methods often rely on prior knowledge of specific dataset biases, which is challenging to acquire a priori. We propose RAZOR (Rewriting And Zero-bias Optimization Refinement), a novel, unsupervised, and data-focused debiasing approach based on text rewriting for shortcut mitigation. RAZOR leverages LLMs to iteratively rewrite potentially biased text segments by replacing them with heuristically selected alternatives in a shortcut space defined by token statistics and positional information. This process aims to align surface-level text features more closely with diverse label distributions, thereby promoting the learning of genuine linguistic patterns. Compared with unsupervised SoTA models, RAZOR improves by 3.5% on the FEVER and 6.5% on MNLI and SNLI datasets according to the F1 score. Additionally, RAZOR effectively mitigates specific known biases, reducing bias-related terms by x2 without requiring prior bias information, a result that is on par with SoTA models that leverage prior information. Our work prioritizes data manipulation over architectural modifications, emphasizing the pivotal role of data quality in enhancing model performance and fairness. This research contributes to developing more robust evaluation benchmarks for debiasing methods by incorporating metrics for bias reduction and overall model efficacy.
comment: Shuo and Bardh contributed equally. Accepted to AAAI'25
☆ FlexLLM: Exploring LLM Customization for Moving Target Defense on Black-Box LLMs Against Jailbreak Attacks
Defense in large language models (LLMs) is crucial to counter the numerous attackers exploiting these systems to generate harmful content through manipulated prompts, known as jailbreak attacks. Although many defense strategies have been proposed, they often require access to the model's internal structure or need additional training, which is impractical for service providers using LLM APIs, such as OpenAI APIs or Claude APIs. In this paper, we propose a moving target defense approach that alters decoding hyperparameters to enhance model robustness against various jailbreak attacks. Our approach does not require access to the model's internal structure and incurs no additional training costs. The proposed defense includes two key components: (1) optimizing the decoding strategy by identifying and adjusting decoding hyperparameters that influence token generation probabilities, and (2) transforming the decoding hyperparameters and model system prompts into dynamic targets, which are continuously altered during each runtime. By continuously modifying decoding strategies and prompts, the defense effectively mitigates the existing attacks. Our results demonstrate that our defense is the most effective against jailbreak attacks in three of the models tested when using LLMs as black-box APIs. Moreover, our defense offers lower inference costs and maintains comparable response quality, making it a potential layer of protection when used alongside other defense methods.
☆ Searching for Structure: Investigating Emergent Communication with Large Language Models
Human languages have evolved to be structured through repeated language learning and use. These processes introduce biases that operate during language acquisition and shape linguistic systems toward communicative efficiency. In this paper, we investigate whether the same happens if artificial languages are optimised for implicit biases of Large Language Models (LLMs). To this end, we simulate a classical referential game in which LLMs learn and use artificial languages. Our results show that initially unstructured holistic languages are indeed shaped to have some structural properties that allow two LLM agents to communicate successfully. Similar to observations in human experiments, generational transmission increases the learnability of languages, but can at the same time result in non-humanlike degenerate vocabularies. Taken together, this work extends experimental findings, shows that LLMs can be used as tools in simulations of language evolution, and opens possibilities for future human-machine experiments in this field.
☆ ChocoLlama: Lessons Learned From Teaching Llamas Dutch
While Large Language Models (LLMs) have shown remarkable capabilities in natural language understanding and generation, their performance often lags in lower-resource, non-English languages due to biases in the training data. In this work, we explore strategies for adapting the primarily English LLMs (Llama-2 and Llama-3) to Dutch, a language spoken by 30 million people worldwide yet often underrepresented in LLM development. We collect 104GB of Dutch text ($32$B tokens) from various sources to first apply continued pretraining using low-rank adaptation (LoRA), complemented with Dutch posttraining strategies provided by prior work. For Llama-2, we consider using (i) the tokenizer of the original model, and (ii) training a new, Dutch-specific tokenizer combined with embedding reinitialization. We evaluate our adapted models, ChocoLlama-2, both on standard benchmarks and a novel Dutch benchmark, ChocoLlama-Bench. Our results demonstrate that LoRA can effectively scale for language adaptation, and that tokenizer modification with careful weight reinitialization can improve performance. Notably, Llama-3 was released during the course of this project and, upon evaluation, demonstrated superior Dutch capabilities compared to our Dutch-adapted versions of Llama-2. We hence apply the same adaptation technique to Llama-3, using its original tokenizer. While our adaptation methods enhanced Llama-2's Dutch capabilities, we found limited gains when applying the same techniques to Llama-3. This suggests that for ever improving, multilingual foundation models, language adaptation techniques may benefit more from focusing on language-specific posttraining rather than on continued pretraining. We hope this work contributes to the broader understanding of adapting LLMs to lower-resource languages, and to the development of Dutch LLMs in particular.
☆ Piece of Table: A Divide-and-Conquer Approach for Selecting Sub-Tables in Table Question Answering
Applying language models (LMs) to tables is challenging due to the inherent structural differences between two-dimensional tables and one-dimensional text for which the LMs were originally designed. Furthermore, when applying linearized tables to LMs, the maximum token lengths often imposed in self-attention calculations make it difficult to comprehensively understand the context spread across large tables. To address these challenges, we present PieTa (Piece of Table), a new framework for sub-table-based question answering (QA). PieTa operates through an iterative process of dividing tables into smaller windows, using LMs to select relevant cells within each window, and merging these cells into a sub-table. This multi-resolution approach captures dependencies across multiple rows and columns while avoiding the limitations caused by long context inputs. Instantiated as a simple iterative sub-table union algorithm, PieTa demonstrates improved performance over previous sub-table-based QA approaches.
☆ DRUM: Learning Demonstration Retriever for Large MUlti-modal Models
Recently, large language models (LLMs) have demonstrated impressive capabilities in dealing with new tasks with the help of in-context learning (ICL). In the study of Large Vision-Language Models (LVLMs), when implementing ICL, researchers usually adopts the naive strategies like fixed demonstrations across different samples, or selecting demonstrations directly via a visual-language embedding model. These methods does not guarantee the configured demonstrations fit the need of the LVLMs. To address this issue, we now propose a novel framework, \underline{d}emonstration \underline{r}etriever for large m\underline{u}lti-modal \underline{m}odel (DRUM), which fine-tunes the visual-language embedding model to better meet the LVLM's needs. First, we discuss the retrieval strategies for a visual-language task, assuming an embedding model is given. And we propose to concate the image and text embeddings to enhance the retrieval performance. Second, we propose to re-rank the demonstrations retrieved by the embedding model via the LVLM's feedbacks, and calculate a list-wise ranking loss for training the embedding model. Third, we propose an iterative demonstration mining strategy to improve the training of the embedding model. Through extensive experiments on 3 types of visual-language tasks, 7 benchmark datasets, our DRUM framework is proven to be effective in boosting the LVLM's in-context learning performance via retrieving more proper demonstrations.
☆ Adapting to Non-Stationary Environments: Multi-Armed Bandit Enhanced Retrieval-Augmented Generation on Knowledge Graphs AAAI 2025
Despite the superior performance of Large language models on many NLP tasks, they still face significant limitations in memorizing extensive world knowledge. Recent studies have demonstrated that leveraging the Retrieval-Augmented Generation (RAG) framework, combined with Knowledge Graphs that encapsulate extensive factual data in a structured format, robustly enhances the reasoning capabilities of LLMs. However, deploying such systems in real-world scenarios presents challenges: the continuous evolution of non-stationary environments may lead to performance degradation and user satisfaction requires a careful balance of performance and responsiveness. To address these challenges, we introduce a Multi-objective Multi-Armed Bandit enhanced RAG framework, supported by multiple retrieval methods with diverse capabilities under rich and evolving retrieval contexts in practice. Within this framework, each retrieval method is treated as a distinct ``arm''. The system utilizes real-time user feedback to adapt to dynamic environments, by selecting the appropriate retrieval method based on input queries and the historical multi-objective performance of each arm. Extensive experiments conducted on two benchmark KGQA datasets demonstrate that our method significantly outperforms baseline methods in non-stationary settings while achieving state-of-the-art performance in stationary environments. Code and data are available at https://github.com/FUTUREEEEEE/Dynamic-RAG.git
comment: AAAI 2025
☆ SST framework for Document Matching
Long-form document matching aims to judge the relevance between two documents and has been applied to various scenarios. Most existing works utilize hierarchical or long context models to process documents, which achieve coarse understanding but may ignore details. Some researchers construct a document view with similar sentences about aligned document subtopics to focus on detailed matching signals. However, a long document generally contains multiple subtopics. The matching signals are heterogeneous from multiple topics. Considering only the homologous aligned subtopics may not be representative enough and may cause biased modeling. In this paper, we introduce a new framework to model representative matching signals. First, we propose to capture various matching signals through subtopics of document pairs. Next, We construct multiple document views based on subtopics to cover heterogeneous and valuable details. However, existing spatial aggregation methods like attention, which integrate all these views simultaneously, are hard to integrate heterogeneous information. Instead, we propose temporal aggregation, which effectively integrates different views gradually as the training progresses. Experimental results show that our learning framework is effective on several document-matching tasks, including news duplication and legal case retrieval.
☆ CoPrUS: Consistency Preserving Utterance Synthesis towards more realistic benchmark dialogues COLING 2025
Large-scale Wizard-Of-Oz dialogue datasets have enabled the training of deep learning-based dialogue systems. While they are successful as benchmark datasets, they lack certain types of utterances, which would make them more realistic. In this work, we investigate the creation of synthetic communication errors in an automatic pipeline. Based on linguistic theory, we propose and follow a simple error taxonomy. We focus on three types of miscommunications that could happen in real-world dialogues but are underrepresented in the benchmark dataset: misunderstandings, non-understandings and vaguely related questions. Our two-step approach uses a state-of-the-art Large Language Model (LLM) to first create the error and secondly the repairing utterance. We perform Language Model-based evaluation to ensure the quality of the generated utterances. We apply the method to the MultiWOZ dataset and evaluate it both qualitatively and empirically as well as with human judges. Our results indicate that current LLMs can aid in adding post-hoc miscommunications to benchmark datasets as a form of data augmentation. We publish the resulting dataset, in which nearly 1900 dialogues have been modified, as CoPrUS-MultiWOZ to facilitate future work on dialogue systems.
comment: Accepted at COLING 2025 (main, long paper)
☆ Bilingual BSARD: Extending Statutory Article Retrieval to Dutch COLING
Statutory article retrieval plays a crucial role in making legal information more accessible to both laypeople and legal professionals. Multilingual countries like Belgium present unique challenges for retrieval models due to the need for handling legal issues in multiple languages. Building on the Belgian Statutory Article Retrieval Dataset (BSARD) in French, we introduce the bilingual version of this dataset, bBSARD. The dataset contains parallel Belgian statutory articles in both French and Dutch, along with legal questions from BSARD and their Dutch translation. Using bBSARD, we conduct extensive benchmarking of retrieval models available for Dutch and French. Our benchmarking setup includes lexical models, zero-shot dense models, and fine-tuned small foundation models. Our experiments show that BM25 remains a competitive baseline compared to many zero-shot dense models in both languages. We also observe that while proprietary models outperform open alternatives in the zero-shot setting, they can be matched or surpassed by fine-tuning small language-specific models. Our dataset and evaluation code are publicly available.
comment: To be presented at RegNLP-2025 (COLING)
☆ Causal World Representation in the GPT Model NeurIPS 2024
Are generative pre-trained transformer (GPT) models only trained to predict the next token, or do they implicitly learn a world model from which a sequence is generated one token at a time? We examine this question by deriving a causal interpretation of the attention mechanism in GPT, and suggesting a causal world model that arises from this interpretation. Furthermore, we propose that GPT-models, at inference time, can be utilized for zero-shot causal structure learning for in-distribution sequences. Empirical evaluation is conducted in a controlled synthetic environment using the setup and rules of the Othello board game. A GPT, pre-trained on real-world games played with the intention of winning, is tested on synthetic data that only adheres to the game rules. We find that the GPT model tends to generate next moves that adhere to the game rules for sequences for which the attention mechanism encodes a causal structure with high confidence. In general, in cases for which the GPT model generates moves that do not adhere to the game rules, it also fails to capture any causal structure.
comment: NeurIPS 2024 Workshop on Causality and Large Models (CaLM)
☆ Knowledge Graph Guided Evaluation of Abstention Techniques
To deploy language models safely, it is crucial that they abstain from responding to inappropriate requests. Several prior studies test the safety promises of models based on their effectiveness in blocking malicious requests. In this work, we focus on evaluating the underlying techniques that cause models to abstain. We create SELECT, a benchmark derived from a set of benign concepts (e.g., "rivers") from a knowledge graph. The nature of SELECT enables us to isolate the effects of abstention techniques from other safety training procedures, as well as evaluate their generalization and specificity. Using SELECT, we benchmark different abstention techniques over six open-weight and closed-source models. We find that the examined techniques indeed cause models to abstain with over $80\%$ abstention rates. However, these techniques are not as effective for descendants of the target concepts, with refusal rates declining by $19\%$. We also characterize the generalization-vs-specificity trade-offs for different techniques. Overall, no single technique is invariably better than the others. Our findings call for a careful evaluation of different aspects of abstention, and hopefully inform practitioners of various trade-offs involved.
☆ Optimizing Alignment with Less: Leveraging Data Augmentation for Personalized Evaluation
Automatic evaluation by large language models (LLMs) is a prominent topic today; however, judgment and evaluation tasks are often subjective and influenced by various factors, making adaptation challenging. While many studies demonstrate the capabilities of state-of-the-art proprietary LLMs in comparison to human evaluators, they often struggle to adapt to reference evaluators over time, a requirement for achieving personalized judgment. Additionally, numerous works have attempted to apply open LLMs as judges or evaluators, but these efforts frequently overlook the limitations of working with scarce data. Personalized judgment is inherently associated with limited data scenarios, which are common in many real-world problems. Our work aims to present a data augmentation technique to select a more effective sample from limited data in order to align an open LLM with human preference. Our work achieves approximately 7% improvements in Pearson correlation with a reference judge over the baseline,and 30% improvement over the base model (Llama3.1-8B-Instruct) in the mathematical reasoning evaluation task. demonstrating that augmenting selecting more effective preference data enables our approach to surpass baseline methods.
☆ RAG-based Question Answering over Heterogeneous Data and Text
This article presents the QUASAR system for question answering over unstructured text, structured tables, and knowledge graphs, with unified treatment of all sources. The system adopts a RAG-based architecture, with a pipeline of evidence retrieval followed by answer generation, with the latter powered by a moderate-sized language model. Additionally and uniquely, QUASAR has components for question understanding, to derive crisper input for evidence retrieval, and for re-ranking and filtering the retrieved evidence before feeding the most informative pieces into the answer generation. Experiments with three different benchmarks demonstrate the high answering quality of our approach, being on par with or better than large GPT models, while keeping the computational cost and energy consumption orders of magnitude lower.
comment: IEEE Data Engineering Bulletin -- December 2024 Edition on RAG
☆ Composing or Not Composing? Towards Distributional Construction Grammars
The mechanisms of comprehension during language processing remains an open question. Classically, building the meaning of a linguistic utterance is said to be incremental, step-by-step, based on a compositional process. However, many different works have shown for a long time that non-compositional phenomena are also at work. It is therefore necessary to propose a framework bringing together both approaches. We present in this paper an approach based on Construction Grammars and completing this framework in order to account for these different mechanisms. We propose first a formal definition of this framework by completing the feature structure representation proposed in Sign-Based Construction Grammars. In a second step, we present a general representation of the meaning based on the interaction of constructions, frames and events. This framework opens the door to a processing mechanism for building the meaning based on the notion of activation evaluated in terms of similarity and unification. This new approach integrates features from distributional semantics into the constructionist framework, leading to what we call Distributional Construction Grammars.
☆ Generating Knowledge Graphs from Large Language Models: A Comparative Study of GPT-4, LLaMA 2, and BERT
Knowledge Graphs (KGs) are essential for the functionality of GraphRAGs, a form of Retrieval-Augmented Generative Systems (RAGs) that excel in tasks requiring structured reasoning and semantic understanding. However, creating KGs for GraphRAGs remains a significant challenge due to accuracy and scalability limitations of traditional methods. This paper introduces a novel approach leveraging large language models (LLMs) like GPT-4, LLaMA 2 (13B), and BERT to generate KGs directly from unstructured data, bypassing traditional pipelines. Using metrics such as Precision, Recall, F1-Score, Graph Edit Distance, and Semantic Similarity, we evaluate the models' ability to generate high-quality KGs. Results demonstrate that GPT-4 achieves superior semantic fidelity and structural accuracy, LLaMA 2 excels in lightweight, domain-specific graphs, and BERT provides insights into challenges in entity-relationship modeling. This study underscores the potential of LLMs to streamline KG creation and enhance GraphRAG accessibility for real-world applications, while setting a foundation for future advancements.
comment: 4 pages, 4 figures, 3 tables
☆ CMT: A Memory Compression Method for Continual Knowledge Learning of Large Language Models AAAI 2025
Large Language Models (LLMs) need to adapt to the continuous changes in data, tasks, and user preferences. Due to their massive size and the high costs associated with training, LLMs are not suitable for frequent retraining. However, updates are necessary to keep them in sync with rapidly evolving human knowledge. To address these challenges, this paper proposes the Compression Memory Training (CMT) method, an efficient and effective online adaptation framework for LLMs that features robust knowledge retention capabilities. Inspired by human memory mechanisms, CMT compresses and extracts information from new documents to be stored in a memory bank. When answering to queries related to these new documents, the model aggregates these document memories from the memory bank to better answer user questions. The parameters of the LLM itself do not change during training and inference, reducing the risk of catastrophic forgetting. To enhance the encoding, retrieval, and aggregation of memory, we further propose three new general and flexible techniques, including memory-aware objective, self-matching and top-aggregation. Extensive experiments conducted on three continual learning datasets (i.e., StreamingQA, SQuAD and ArchivalQA) demonstrate that the proposed method improves model adaptability and robustness across multiple base LLMs (e.g., +4.07 EM & +4.19 F1 in StreamingQA with Llama-2-7b).
comment: AAAI 2025; Pre-print
☆ A Review of Challenges in Speech-based Conversational AI for Elderly Care
Artificially intelligent systems optimized for speech conversation are appearing at a fast pace. Such models are interesting from a healthcare perspective, as these voice-controlled assistants may support the elderly and enable remote health monitoring. The bottleneck for efficacy, however, is how well these devices work in practice and how the elderly experience them, but research on this topic is scant. We review elderly use of voice-controlled AI and highlight various user- and technology-centered issues, that need to be considered before effective speech-controlled AI for elderly care can be realized.
comment: Accepted for publication at Medical Informatics Europe 2025 conference, Glasgow. 5 pages, 1 figure, 1 table
☆ Algorithmic Phase Transitions in Language Models: A Mechanistic Case Study of Arithmetic
Zero-shot capabilities of large language models make them powerful tools for solving a range of tasks without explicit training. It remains unclear, however, how these models achieve such performance, or why they can zero-shot some tasks but not others. In this paper, we shed some light on this phenomenon by defining and investigating algorithmic stability in language models -- changes in problem-solving strategy employed by the model as a result of changes in task specification. We focus on a task where algorithmic stability is needed for generalization: two-operand arithmetic. Surprisingly, we find that Gemma-2-2b employs substantially different computational models on closely related subtasks, i.e. four-digit versus eight-digit addition. Our findings suggest that algorithmic instability may be a contributing factor to language models' poor zero-shot performance across certain logical reasoning tasks, as they struggle to abstract different problem-solving strategies and smoothly transition between them.
comment: 10 pages, 5 figures
☆ SpecFuse: Ensembling Large Language Models via Next-Segment Prediction
Ensembles of generative large language models (LLMs) can integrate the strengths of different LLMs to compensate for the limitations of individual models. However, recent work has focused on training an additional fusion model to combine complete responses from multiple LLMs, failing to tap into their collaborative potential to generate higher-quality responses. Moreover, as the additional fusion model is trained on a specialized dataset, these methods struggle with generalizing to open-domain queries from online users. In this paper, we propose SpecFuse, a novel ensemble framework that outputs the fused result by iteratively producing the next segment through collaboration among LLMs. This is achieved through cyclic execution of its inference and verification components. In each round, the inference component invokes each base LLM to generate candidate segments in parallel, and the verify component calls these LLMs again to predict the ranking of the segments. The top-ranked segment is then broadcast to all LLMs, encouraging them to generate higher-quality segments in the next round. This approach also allows the base LLMs to be plug-and-play, without any training or adaptation, avoiding generalization limitations. Furthermore, to conserve computational resources, we propose a model exit mechanism that dynamically excludes models exhibiting poor performance in previous rounds during each query response. In this way, it effectively reduces the number of model calls while maintaining overall performance.
comment: 15 pages, 5 figures
☆ My Words Imply Your Opinion: Reader Agent-Based Propagation Enhancement for Personalized Implicit Emotion Analysis
In implicit emotion analysis (IEA), the subtlety of emotional expressions makes it particularly sensitive to user-specific characteristics. Existing studies often inject personalization into the analysis by focusing on the authorial dimension of the emotional text. However, these methods overlook the potential influence of the intended reader on the reaction of implicit emotions. In this paper, we refine the IEA task to Personalized Implicit Emotion Analysis (PIEA) and introduce the RAPPIE model, a novel framework designed to address the issue of missing user information within this task. In particular, 1) we create reader agents based on the Large Language Model to simulate reader reactions, to address challenges of the spiral of silence and data incompleteness encountered when acquiring reader feedback information. 2) We establish a reader propagation role system and develop a role-aware emotion propagation multi-view graph learning model, which effectively deals with the sparsity of reader information by utilizing the distribution of propagation roles. 3) We annotate two Chinese PIEA datasets with detailed user metadata, thereby addressing the limitation of prior datasets that primarily focus on textual content annotation. Extensive experiments on these datasets indicate that the RAPPIE model outperforms current state-of-the-art baselines, highlighting the significance and efficacy of incorporating reader feedback into the PIEA process.
☆ Towards Predictive Communication with Brain-Computer Interfaces integrating Large Language Models
This perspective article aims at providing an outline of the state of the art and future developments towards the integration of cutting-edge predictive language models with BCI. A synthetic overview of early and more recent linguistic models, from natural language processing (NLP) models to recent LLM, that to a varying extent improved predictive writing systems, is first provided. Second, a summary of previous BCI implementations integrating language models is presented. The few preliminary studies investigating the possible combination of LLM with BCI spellers to efficiently support fast communication and control are then described. Finally, current challenges and limitations towards the full integration of LLM with BCI systems are discussed. Recent investigations suggest that the combination of LLM with BCI might drastically improve human-computer interaction in patients with motor or language disorders as well as in healthy individuals. In particular, the pretrained autoregressive transformer models, such as GPT, that capitalize from parallelization, learning through pre-training and fine-tuning, promise a substantial improvement of BCI for communication with respect to previous systems incorporating simpler language models. Indeed, among various models, the GPT-2 was shown to represent an excellent candidate for its integration into BCI although testing was only perfomed on simulated conversations and not on real BCI scenarios. Prospectively, the full integration of LLM with advanced BCI systems might lead to a big leap forward towards fast, efficient and user-adaptive neurotechnology.
☆ Frame Representation Hypothesis: Multi-Token LLM Interpretability and Concept-Guided Text Generation
Interpretability is a key challenge in fostering trust for Large Language Models (LLMs), which stems from the complexity of extracting reasoning from model's parameters. We present the Frame Representation Hypothesis, a theoretically robust framework grounded in the Linear Representation Hypothesis (LRH) to interpret and control LLMs by modeling multi-token words. Prior research explored LRH to connect LLM representations with linguistic concepts, but was limited to single token analysis. As most words are composed of several tokens, we extend LRH to multi-token words, thereby enabling usage on any textual data with thousands of concepts. To this end, we propose words can be interpreted as frames, ordered sequences of vectors that better capture token-word relationships. Then, concepts can be represented as the average of word frames sharing a common concept. We showcase these tools through Top-k Concept-Guided Decoding, which can intuitively steer text generation using concepts of choice. We verify said ideas on Llama 3.1, Gemma 2, and Phi 3 families, demonstrating gender and language biases, exposing harmful content, but also potential to remediate them, leading to safer and more transparent LLMs. Code is available at https://github.com/phvv-me/frame-representation-hypothesis.git
☆ Filipino Benchmarks for Measuring Sexist and Homophobic Bias in Multilingual Language Models from Southeast Asia COLING 2025
Bias studies on multilingual models confirm the presence of gender-related stereotypes in masked models processing languages with high NLP resources. We expand on this line of research by introducing Filipino CrowS-Pairs and Filipino WinoQueer: benchmarks that assess both sexist and anti-queer biases in pretrained language models (PLMs) handling texts in Filipino, a low-resource language from the Philippines. The benchmarks consist of 7,074 new challenge pairs resulting from our cultural adaptation of English bias evaluation datasets, a process that we document in detail to guide similar forthcoming efforts. We apply the Filipino benchmarks on masked and causal multilingual models, including those pretrained on Southeast Asian data, and find that they contain considerable amounts of bias. We also find that for multilingual models, the extent of bias learned for a particular language is influenced by how much pretraining data in that language a model was exposed to. Our benchmarks and insights can serve as a foundation for future work analyzing and mitigating bias in multilingual models.
comment: Accepted for presentation at The First Workshop on Language Models for Low-Resource Languages (LoResLM) at The 31st International Conference on Computational Linguistics (COLING 2025)
☆ The Rise and Down of Babel Tower: Investigating the Evolution Process of Multilingual Code Large Language Model
Large language models (LLMs) have shown significant multilingual capabilities. However, the mechanisms underlying the development of these capabilities during pre-training are not well understood. In this paper, we use code LLMs as an experimental platform to explore the evolution of multilingual capabilities in LLMs during the pre-training process. Based on our observations, we propose the Babel Tower Hypothesis, which describes the entire process of LLMs acquiring new language capabilities. During the learning process, multiple languages initially share a single knowledge system dominated by the primary language and gradually develop language-specific knowledge systems. We then validate the above hypothesis by tracking the internal states of the LLMs through identifying working languages and language transferring neurons. Experimental results show that the internal state changes of the LLM are consistent with our Babel Tower Hypothesis. Building on these insights, we propose a novel method to construct an optimized pre-training corpus for multilingual code LLMs, which significantly outperforms LLMs trained on the original corpus. The proposed Babel Tower Hypothesis provides new insights into designing pre-training data distributions to achieve optimal multilingual capabilities in LLMs.
☆ Multimodal Sentiment Analysis Based on Causal Reasoning
With the rapid development of multimedia, the shift from unimodal textual sentiment analysis to multimodal image-text sentiment analysis has obtained academic and industrial attention in recent years. However, multimodal sentiment analysis is affected by unimodal data bias, e.g., text sentiment is misleading due to explicit sentiment semantic, leading to low accuracy in the final sentiment classification. In this paper, we propose a novel CounterFactual Multimodal Sentiment Analysis framework (CF-MSA) using causal counterfactual inference to construct multimodal sentiment causal inference. CF-MSA mitigates the direct effect from unimodal bias and ensures heterogeneity across modalities by differentiating the treatment variables between modalities. In addition, considering the information complementarity and bias differences between modalities, we propose a new optimisation objective to effectively integrate different modalities and reduce the inherent bias from each modality. Experimental results on two public datasets, MVSA-Single and MVSA-Multiple, demonstrate that the proposed CF-MSA has superior debiasing capability and achieves new state-of-the-art performances. We will release the code and datasets to facilitate future research.
☆ Enhancing Relation Extraction via Supervised Rationale Verification and Feedback AAAI 2025
Despite the rapid progress that existing automated feedback methods have made in correcting the output of large language models (LLMs), these methods cannot be well applied to the relation extraction (RE) task due to their designated feedback objectives and correction manner. To address this problem, we propose a novel automated feedback framework for RE, which presents a rationale supervisor to verify the rationale and provide re-selected demonstrations as feedback to correct the initial prediction. Specifically, we first design a causal intervention and observation method for to collect biased/unbiased rationales for contrastive training the rationale supervisor. Then, we present a verification-feedback-correction procedure to iteratively enhance LLMs' capability of handling the RE task. Extensive experiments prove that our proposed framework significantly outperforms existing methods.
comment: Accepted to AAAI 2025, camera ready version
☆ HARP: Hesitation-Aware Reframing in Transformer Inference Pass
This paper aims to improve the performance of large language models by addressing the variable computational demands in inference steps, where some tokens require more computational resources than others. We present HARP, a simple modification to "off-the-shelf" Transformer forward pass. Drawing from hesitation and the framing effect in decision-making, HARP selectively applies additional computation when the model encounters uncertainty during token generation. Our method mimics human cognitive processes by pausing at difficult decision points and reframing inputs for a different perspective. Unlike other approaches, HARP is model-agnostic, training-free, and easy to implement. We thoroughly evaluate our method across various downstream tasks and model sizes, demonstrating performance improvements up to +5.16%. Notably, HARP achieves these gains while maintaining inference times twice faster than beam search. Simple and yet with significant gains, HARP offers a practical solution for enhancing the performance of Transformer-based language models with minimal computational impact.
☆ Label-Confidence-Aware Uncertainty Estimation in Natural Language Generation
Large Language Models (LLMs) display formidable capabilities in generative tasks but also pose potential risks due to their tendency to generate hallucinatory responses. Uncertainty Quantification (UQ), the evaluation of model output reliability, is crucial for ensuring the safety and robustness of AI systems. Recent studies have concentrated on model uncertainty by analyzing the relationship between output entropy under various sampling conditions and the corresponding labels. However, these methods primarily focus on measuring model entropy with precision to capture response characteristics, often neglecting the uncertainties associated with greedy decoding results-the sources of model labels, which can lead to biased classification outcomes. In this paper, we explore the biases introduced by greedy decoding and propose a label-confidence-aware (LCA) uncertainty estimation based on Kullback-Leibler (KL) divergence bridging between samples and label source, thus enhancing the reliability and stability of uncertainty assessments. Our empirical evaluations across a range of popular LLMs and NLP datasets reveal that different label sources can indeed affect classification, and that our approach can effectively capture differences in sampling results and label sources, demonstrating more effective uncertainty estimation.
☆ KULTURE Bench: A Benchmark for Assessing Language Model in Korean Cultural Context
Large language models have exhibited significant enhancements in performance across various tasks. However, the complexity of their evaluation increases as these models generate more fluent and coherent content. Current multilingual benchmarks often use translated English versions, which may incorporate Western cultural biases that do not accurately assess other languages and cultures. To address this research gap, we introduce KULTURE Bench, an evaluation framework specifically designed for Korean culture that features datasets of cultural news, idioms, and poetry. It is designed to assess language models' cultural comprehension and reasoning capabilities at the word, sentence, and paragraph levels. Using the KULTURE Bench, we assessed the capabilities of models trained with different language corpora and analyzed the results comprehensively. The results show that there is still significant room for improvement in the models' understanding of texts related to the deeper aspects of Korean culture.
comment: Accepted by the 38th Pacific Asia Conference on Language, Information and Computation
☆ Filling Memory Gaps: Enhancing Continual Semantic Parsing via SQL Syntax Variance-Guided LLMs without Real Data Replay
Continual Semantic Parsing (CSP) aims to train parsers to convert natural language questions into SQL across tasks with limited annotated examples, adapting to the real-world scenario of dynamically updated databases. Previous studies mitigate this challenge by replaying historical data or employing parameter-efficient tuning (PET), but they often violate data privacy or rely on ideal continual learning settings. To address these problems, we propose a new Large Language Model (LLM)-Enhanced Continuous Semantic Parsing method, named LECSP, which alleviates forgetting while encouraging generalization, without requiring real data replay or ideal settings. Specifically, it first analyzes the commonalities and differences between tasks from the SQL syntax perspective to guide LLMs in reconstructing key memories and improving memory accuracy through a calibration strategy. Then, it uses a task-aware dual-teacher distillation framework to promote the accumulation and transfer of knowledge during sequential training. Experimental results on two CSP benchmarks show that our method significantly outperforms existing methods, even those utilizing data replay or ideal settings. Additionally, we achieve generalization performance beyond the upper limits, better adapting to unseen tasks.
☆ Speaker effects in spoken language comprehension
The identity of a speaker significantly influences spoken language comprehension by affecting both perception and expectation. This review explores speaker effects, focusing on how speaker information impacts language processing. We propose an integrative model featuring the interplay between bottom-up perception-based processes driven by acoustic details and top-down expectation-based processes driven by a speaker model. The acoustic details influence lower-level perception, while the speaker model modulates both lower-level and higher-level processes such as meaning interpretation and pragmatic inferences. We define speaker-idiosyncrasy and speaker-demographics effects and demonstrate how bottom-up and top-down processes interact at various levels in different scenarios. This framework contributes to psycholinguistic theory by offering a comprehensive account of how speaker information interacts with linguistic content to shape message construction. We suggest that speaker effects can serve as indices of a language learner's proficiency and an individual's characteristics of social cognition. We encourage future research to extend these findings to AI speakers, probing the universality of speaker effects across humans and artificial agents.
comment: 44 pages, 1 figure
☆ Comateformer: Combined Attention Transformer for Semantic Sentence Matching ECAI 2024
The Transformer-based model have made significant strides in semantic matching tasks by capturing connections between phrase pairs. However, to assess the relevance of sentence pairs, it is insufficient to just examine the general similarity between the sentences. It is crucial to also consider the tiny subtleties that differentiate them from each other. Regrettably, attention softmax operations in transformers tend to miss these subtle differences. To this end, in this work, we propose a novel semantic sentence matching model named Combined Attention Network based on Transformer model (Comateformer). In Comateformer model, we design a novel transformer-based quasi-attention mechanism with compositional properties. Unlike traditional attention mechanisms that merely adjust the weights of input tokens, our proposed method learns how to combine, subtract, or resize specific vectors when building a representation. Moreover, our proposed approach builds on the intuition of similarity and dissimilarity (negative affinity) when calculating dual affinity scores. This allows for a more meaningful representation of relationships between sentences. To evaluate the performance of our proposed model, we conducted extensive experiments on ten public real-world datasets and robustness testing. Experimental results show that our method achieves consistent improvements.
comment: This paper is accepted by 27th EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE (ECAI 2024)
☆ MAPLE: A Framework for Active Preference Learning Guided by Large Language Models
The advent of large language models (LLMs) has sparked significant interest in using natural language for preference learning. However, existing methods often suffer from high computational burdens, taxing human supervision, and lack of interpretability. To address these issues, we introduce MAPLE, a framework for large language model-guided Bayesian active preference learning. MAPLE leverages LLMs to model the distribution over preference functions, conditioning it on both natural language feedback and conventional preference learning feedback, such as pairwise trajectory rankings. MAPLE also employs active learning to systematically reduce uncertainty in this distribution and incorporates a language-conditioned active query selection mechanism to identify informative and easy-to-answer queries, thus reducing human burden. We evaluate MAPLE's sample efficiency and preference inference quality across two benchmarks, including a real-world vehicle route planning benchmark using OpenStreetMap data. Our results demonstrate that MAPLE accelerates the learning process and effectively improves humans' ability to answer queries.
☆ A Review on the Applications of Transformer-based language models for Nucleotide Sequence Analysis
In recent times, Transformer-based language models are making quite an impact in the field of natural language processing. As relevant parallels can be drawn between biological sequences and natural languages, the models used in NLP can be easily extended and adapted for various applications in bioinformatics. In this regard, this paper introduces the major developments of Transformer-based models in the recent past in the context of nucleotide sequences. We have reviewed and analysed a large number of application-based papers on this subject, giving evidence of the main characterizing features and to different approaches that may be adopted to customize such powerful computational machines. We have also provided a structured description of the functioning of Transformers, that may enable even first time users to grab the essence of such complex architectures. We believe this review will help the scientific community in understanding the various applications of Transformer-based language models to nucleotide sequences. This work will motivate the readers to build on these methodologies to tackle also various other problems in the field of bioinformatics.
☆ Modifying AI, Enhancing Essays: How Active Engagement with Generative AI Boosts Writing Quality
Students are increasingly relying on Generative AI (GAI) to support their writing-a key pedagogical practice in education. In GAI-assisted writing, students can delegate core cognitive tasks (e.g., generating ideas and turning them into sentences) to GAI while still producing high-quality essays. This creates new challenges for teachers in assessing and supporting student learning, as they often lack insight into whether students are engaging in meaningful cognitive processes during writing or how much of the essay's quality can be attributed to those processes. This study aimed to help teachers better assess and support student learning in GAI-assisted writing by examining how different writing behaviors, especially those indicative of meaningful learning versus those that are not, impact essay quality. Using a dataset of 1,445 GAI-assisted writing sessions, we applied the cutting-edge method, X-Learner, to quantify the causal impact of three GAI-assisted writing behavioral patterns (i.e., seeking suggestions but not accepting them, seeking suggestions and accepting them as they are, and seeking suggestions and accepting them with modification) on four measures of essay quality (i.e., lexical sophistication, syntactic complexity, text cohesion, and linguistic bias). Our analysis showed that writers who frequently modified GAI-generated text-suggesting active engagement in higher-order cognitive processes-consistently improved the quality of their essays in terms of lexical sophistication, syntactic complexity, and text cohesion. In contrast, those who often accepted GAI-generated text without changes, primarily engaging in lower-order processes, saw a decrease in essay quality. Additionally, while human writers tend to introduce linguistic bias when writing independently, incorporating GAI-generated text-even without modification-can help mitigate this bias.
☆ PrisonBreak: Jailbreaking Large Language Models with Fewer Than Twenty-Five Targeted Bit-flips
We introduce a new class of attacks on commercial-scale (human-aligned) language models that induce jailbreaking through targeted bitwise corruptions in model parameters. Our adversary can jailbreak billion-parameter language models with fewer than 25 bit-flips in all cases$-$and as few as 5 in some$-$using up to 40$\times$ less bit-flips than existing attacks on computer vision models at least 100$\times$ smaller. Unlike prompt-based jailbreaks, our attack renders these models in memory 'uncensored' at runtime, allowing them to generate harmful responses without any input modifications. Our attack algorithm efficiently identifies target bits to flip, offering up to 20$\times$ more computational efficiency than previous methods. This makes it practical for language models with billions of parameters. We show an end-to-end exploitation of our attack using software-induced fault injection, Rowhammer (RH). Our work examines 56 DRAM RH profiles from DDR4 and LPDDR4X devices with different RH vulnerabilities. We show that our attack can reliably induce jailbreaking in systems similar to those affected by prior bit-flip attacks. Moreover, our approach remains effective even against highly RH-secure systems (e.g., 46$\times$ more secure than previously tested systems). Our analyses further reveal that: (1) models with less post-training alignment require fewer bit flips to jailbreak; (2) certain model components, such as value projection layers, are substantially more vulnerable than others; and (3) our method is mechanistically different than existing jailbreaks. Our findings highlight a pressing, practical threat to the language model ecosystem and underscore the need for research to protect these models from bit-flip attacks.
☆ Breaking the Stage Barrier: A Novel Single-Stage Approach to Long Context Extension for Large Language Models
Recently, Large language models (LLMs) have revolutionized Natural Language Processing (NLP). Pretrained LLMs, due to limited training context size, struggle with handling long token sequences, limiting their performance on various downstream tasks. Current solutions toward long context modeling often employ multi-stage continual pertaining, which progressively increases the effective context length through several continual pretraining stages. However, those approaches require extensive manual tuning and human expertise. In this paper, we introduce a novel single-stage continual pretraining method, Head-Adaptive Rotary Position Encoding (HARPE), to equip LLMs with long context modeling capabilities while simplifying the training process. Our HARPE leverages different Rotary Position Encoding (RoPE) base frequency values across different attention heads and directly trains LLMs on the target context length. Extensive experiments on 4 language modeling benchmarks, including the latest RULER benchmark, demonstrate that HARPE excels in understanding and integrating long-context tasks with single-stage training, matching and even outperforming existing multi-stage methods. Our results highlight that HARPE successfully breaks the stage barrier for training LLMs with long context modeling capabilities.
☆ MM-PoE: Multiple Choice Reasoning via. Process of Elimination using Multi-Modal Models
This paper introduces Multiple Choice Reasoning via. Process of Elimination using Multi-Modal models, herein referred to as Multi-Modal Process of Elimination (MM-PoE). This novel methodology is engineered to augment the efficacy of Vision-Language Models (VLMs) in multiple-choice visual reasoning tasks. Diverging from conventional approaches that evaluate each option independently, MM-PoE employs a dual-step scoring paradigm that initially identifies and excludes implausible choices, subsequently concentrating on the most probable remaining options. This method emulates human test-taking strategies, where individuals typically eliminate clearly incorrect answers prior to selecting the optimal response. Our empirical evaluations, conducted across three benchmark datasets, reveal that MM-PoE significantly improves both zero-shot and few-shot performance of contemporary state-of-the-art VLMs. Critically, this approach not only broadens the application of the elimination process to multi-modal contexts but also allows few-shot experiments, thereby addressing two principal limitations concerning usage of PoE only in zero-shot settings and only with a language-only framework. As a result, MM-PoE not only refines the reasoning capabilities of VLMs but also broadens their applicability to complex visual question-answering scenarios. All code and documentation supporting our work are available at https://pypi.org/project/mm-poe/, enabling researchers and practitioners to easily integrate and further develop these techniques.
☆ Political Actor Agent: Simulating Legislative System for Roll Call Votes Prediction with Large Language Models AAAI 2025
Predicting roll call votes through modeling political actors has emerged as a focus in quantitative political science and computer science. Widely used embedding-based methods generate vectors for legislators from diverse data sets to predict legislative behaviors. However, these methods often contend with challenges such as the need for manually predefined features, reliance on extensive training data, and a lack of interpretability. Achieving more interpretable predictions under flexible conditions remains an unresolved issue. This paper introduces the Political Actor Agent (PAA), a novel agent-based framework that utilizes Large Language Models to overcome these limitations. By employing role-playing architectures and simulating legislative system, PAA provides a scalable and interpretable paradigm for predicting roll-call votes. Our approach not only enhances the accuracy of predictions but also offers multi-view, human-understandable decision reasoning, providing new insights into political actor behaviors. We conducted comprehensive experiments using voting records from the 117-118th U.S. House of Representatives, validating the superior performance and interpretability of PAA. This study not only demonstrates PAA's effectiveness but also its potential in political science research.
comment: Accepted at AAAI 2025
☆ Bridging the Gap for Test-Time Multimodal Sentiment Analysis AAAI 2025
Multimodal sentiment analysis (MSA) is an emerging research topic that aims to understand and recognize human sentiment or emotions through multiple modalities. However, in real-world dynamic scenarios, the distribution of target data is always changing and different from the source data used to train the model, which leads to performance degradation. Common adaptation methods usually need source data, which could pose privacy issues or storage overheads. Therefore, test-time adaptation (TTA) methods are introduced to improve the performance of the model at inference time. Existing TTA methods are always based on probabilistic models and unimodal learning, and thus can not be applied to MSA which is often considered as a multimodal regression task. In this paper, we propose two strategies: Contrastive Adaptation and Stable Pseudo-label generation (CASP) for test-time adaptation for multimodal sentiment analysis. The two strategies deal with the distribution shifts for MSA by enforcing consistency and minimizing empirical risk, respectively. Extensive experiments show that CASP brings significant and consistent improvements to the performance of the model across various distribution shift settings and with different backbones, demonstrating its effectiveness and versatility. Our codes are available at https://github.com/zrguo/CASP.
comment: Accepted to AAAI 2025
☆ A Review of Human Emotion Synthesis Based on Generative Technology
Human emotion synthesis is a crucial aspect of affective computing. It involves using computational methods to mimic and convey human emotions through various modalities, with the goal of enabling more natural and effective human-computer interactions. Recent advancements in generative models, such as Autoencoders, Generative Adversarial Networks, Diffusion Models, Large Language Models, and Sequence-to-Sequence Models, have significantly contributed to the development of this field. However, there is a notable lack of comprehensive reviews in this field. To address this problem, this paper aims to address this gap by providing a thorough and systematic overview of recent advancements in human emotion synthesis based on generative models. Specifically, this review will first present the review methodology, the emotion models involved, the mathematical principles of generative models, and the datasets used. Then, the review covers the application of different generative models to emotion synthesis based on a variety of modalities, including facial images, speech, and text. It also examines mainstream evaluation metrics. Additionally, the review presents some major findings and suggests future research directions, providing a comprehensive understanding of the role of generative technology in the nuanced domain of emotion synthesis.
comment: 25 pages, 10 figures
☆ Exploring Coding Spot: Understanding Parametric Contributions to LLM Coding Performance
Large Language Models (LLMs) have demonstrated notable proficiency in both code generation and comprehension across multiple programming languages. However, the mechanisms underlying this proficiency remain underexplored, particularly with respect to whether distinct programming languages are processed independently or within a shared parametric region. Drawing an analogy to the specialized regions of the brain responsible for distinct cognitive functions, we introduce the concept of Coding Spot, a specialized parametric region within LLMs that facilitates coding capabilities. Our findings identify this Coding Spot and show that targeted modifications to this subset significantly affect performance on coding tasks, while largely preserving non-coding functionalities. This compartmentalization mirrors the functional specialization observed in cognitive neuroscience, where specific brain regions are dedicated to distinct tasks, suggesting that LLMs may similarly employ specialized parameter regions for different knowledge domains.
☆ Maya: An Instruction Finetuned Multilingual Multimodal Model
The rapid development of large Vision-Language Models (VLMs) has led to impressive results on academic benchmarks, primarily in widely spoken languages. However, significant gaps remain in the ability of current VLMs to handle low-resource languages and varied cultural contexts, largely due to a lack of high-quality, diverse, and safety-vetted data. Consequently, these models often struggle to understand low-resource languages and cultural nuances in a manner free from toxicity. To address these limitations, we introduce Maya, an open-source Multimodal Multilingual model. Our contributions are threefold: 1) a multilingual image-text pretraining dataset in eight languages, based on the LLaVA pretraining dataset; 2) a thorough analysis of toxicity within the LLaVA dataset, followed by the creation of a novel toxicity-free version across eight languages; and 3) a multilingual image-text model supporting these languages, enhancing cultural and linguistic comprehension in vision-language tasks. Code available at https://github.com/nahidalam/maya.
☆ Predictable Emergent Abilities of LLMs: Proxy Tasks Are All You Need
While scaling laws optimize training configurations for large language models (LLMs) through experiments on smaller or early-stage models, they fail to predict emergent abilities due to the absence of such capabilities in these models. To address this, we propose a method that predicts emergent abilities by leveraging proxy tasks. We begin by establishing relevance metrics between the target task and candidate tasks based on performance differences across multiple models. These candidate tasks are then validated for robustness with small model ensembles, leading to the selection of the most appropriate proxy tasks. The predicted performance on the target task is then derived by integrating the evaluation results of these proxies. In a case study on tool utilization capabilities, our method demonstrated a strong correlation between predicted and actual performance, confirming its effectiveness.
☆ Improving the Natural Language Inference robustness to hard dataset by data augmentation and preprocessing
Natural Language Inference (NLI) is the task of inferring whether the hypothesis can be justified by the given premise. Basically, we classify the hypothesis into three labels(entailment, neutrality and contradiction) given the premise. NLI was well studied by the previous researchers. A number of models, especially the transformer based ones, have achieved significant improvement on these tasks. However, it is reported that these models are suffering when they are dealing with hard datasets. Particularly, they perform much worse when dealing with unseen out-of-distribution premise and hypothesis. They may not understand the semantic content but learn the spurious correlations. In this work, we propose the data augmentation and preprocessing methods to solve the word overlap, numerical reasoning and length mismatch problems. These methods are general methods that do not rely on the distribution of the testing data and they help improve the robustness of the models.
☆ QAPyramid: Fine-grained Evaluation of Content Selection for Text Summarization
How to properly conduct human evaluations for text summarization is a longstanding challenge. The Pyramid human evaluation protocol, which assesses content selection by breaking the reference summary into sub-units and verifying their presence in the system summary, has been widely adopted. However, it suffers from a lack of systematicity in the definition and granularity of the sub-units. We address these problems by proposing QAPyramid, which decomposes each reference summary into finer-grained question-answer (QA) pairs according to the QA-SRL framework. We collect QA-SRL annotations for reference summaries from CNN/DM and evaluate 10 summarization systems, resulting in 8.9K QA-level annotations. We show that, compared to Pyramid, QAPyramid provides more systematic and fine-grained content selection evaluation while maintaining high inter-annotator agreement without needing expert annotations. Furthermore, we propose metrics that automate the evaluation pipeline and achieve higher correlations with QAPyramid than other widely adopted metrics, allowing future work to accurately and efficiently benchmark summarization systems.
comment: The first two authors contributed equally. Code: https://github.com/ZhangShiyue/QAPyramid
☆ Defensive Dual Masking for Robust Adversarial Defense
The field of textual adversarial defenses has gained considerable attention in recent years due to the increasing vulnerability of natural language processing (NLP) models to adversarial attacks, which exploit subtle perturbations in input text to deceive models. This paper introduces the Defensive Dual Masking (DDM) algorithm, a novel approach designed to enhance model robustness against such attacks. DDM utilizes a unique adversarial training strategy where [MASK] tokens are strategically inserted into training samples to prepare the model to handle adversarial perturbations more effectively. During inference, potentially adversarial tokens are dynamically replaced with [MASK] tokens to neutralize potential threats while preserving the core semantics of the input. The theoretical foundation of our approach is explored, demonstrating how the selective masking mechanism strengthens the model's ability to identify and mitigate adversarial manipulations. Our empirical evaluation across a diverse set of benchmark datasets and attack mechanisms consistently shows that DDM outperforms state-of-the-art defense techniques, improving model accuracy and robustness. Moreover, when applied to Large Language Models (LLMs), DDM also enhances their resilience to adversarial attacks, providing a scalable defense mechanism for large-scale NLP applications.
comment: First version
☆ Machines of Meaning
One goal of Artificial Intelligence is to learn meaningful representations for natural language expressions, but what this entails is not always clear. A variety of new linguistic behaviours present themselves embodied as computers, enhanced humans, and collectives with various kinds of integration and communication. But to measure and understand the behaviours generated by such systems, we must clarify the language we use to talk about them. Computational models are often confused with the phenomena they try to model and shallow metaphors are used as justifications for (or to hype) the success of computational techniques on many tasks related to natural language; thus implying their progress toward human-level machine intelligence without ever clarifying what that means. This paper discusses the challenges in the specification of "machines of meaning", machines capable of acquiring meaningful semantics from natural language in order to achieve their goals. We characterize "meaning" in a computational setting, while highlighting the need for detachment from anthropocentrism in the study of the behaviour of machines of meaning. The pressing need to analyse AI risks and ethics requires a proper measurement of its capabilities which cannot be productively studied and explained while using ambiguous language. We propose a view of "meaning" to facilitate the discourse around approaches such as neural language models and help broaden the research perspectives for technology that facilitates dialogues between humans and machines.
☆ HalluCana: Fixing LLM Hallucination with A Canary Lookahead
In this paper, we present HalluCana, a canary lookahead to detect and correct factuality hallucinations of Large Language Models (LLMs) in long-form generation. HalluCana detects and intervenes as soon as traces of hallucination emerge, during and even before generation. To support timely detection, we exploit the internal factuality representation in the LLM hidden space, where we investigate various proxies to the LLMs' factuality self-assessment, and discuss its relation to the models' context familiarity from their pre-training. On biography generation, our method improves generation quality by up to 2.5x, while consuming over 6 times less compute.
☆ Forking Paths in Neural Text Generation
Estimating uncertainty in Large Language Models (LLMs) is important for properly evaluating LLMs, and ensuring safety for users. However, prior approaches to uncertainty estimation focus on the final answer in generated text, ignoring intermediate steps that might dramatically impact the outcome. We hypothesize that there exist key forking tokens, such that re-sampling the system at those specific tokens, but not others, leads to very different outcomes. To test this empirically, we develop a novel approach to representing uncertainty dynamics across individual tokens of text generation, and applying statistical models to test our hypothesis. Our approach is highly flexible: it can be applied to any dataset and any LLM, without fine tuning or accessing model weights. We use our method to analyze LLM responses on 7 different tasks across 4 domains, spanning a wide range of typical use cases. We find many examples of forking tokens, including surprising ones such as punctuation marks, suggesting that LLMs are often just a single token away from saying something very different.
☆ HEDS 3.0: The Human Evaluation Data Sheet Version 3.0
This paper presents version 3.0 of the Human Evaluation Datasheet (HEDS). This update is the result of our experience using HEDS in the context of numerous recent human evaluation experiments, including reproduction studies, and of feedback received. Our main overall goal was to improve clarity, and to enable users to complete the datasheet more consistently and comparably. The HEDS 3.0 package consists of the digital data sheet, documentation, and code for exporting completed data sheets as latex files, all available from the HEDS GitHub.
comment: arXiv admin note: substantial text overlap with arXiv:2103.09710
☆ Style-agnostic evaluation of ASR using multiple reference transcripts
Word error rate (WER) as a metric has a variety of limitations that have plagued the field of speech recognition. Evaluation datasets suffer from varying style, formality, and inherent ambiguity of the transcription task. In this work, we attempt to mitigate some of these differences by performing style-agnostic evaluation of ASR systems using multiple references transcribed under opposing style parameters. As a result, we find that existing WER reports are likely significantly over-estimating the number of contentful errors made by state-of-the-art ASR systems. In addition, we have found our multireference method to be a useful mechanism for comparing the quality of ASR models that differ in the stylistic makeup of their training data and target task.
☆ Asking Again and Again: Exploring LLM Robustness to Repeated Questions
This study examines whether large language models (LLMs), such as ChatGPT, specifically the latest GPT-4o-mini, exhibit sensitivity to repeated prompts and whether repeating a question can improve response accuracy. We hypothesize that reiterating a question within a single prompt might enhance the model's focus on key elements of the query. To test this, we evaluate ChatGPT's performance on a large sample of two reading comprehension datasets under both open-book and closed-book settings, varying the repetition of each question to 1, 3, or 5 times per prompt. Our findings indicate that the model does not demonstrate sensitivity to repeated questions, highlighting its robustness and consistency in this context.
☆ Identifying Quantum Mechanical Statistics in Italian Corpora
We present a theoretical and empirical investigation of the statistical behaviour of the words in a text produced by human language. To this aim, we analyse the word distribution of various texts of Italian language selected from a specific literary corpus. We firstly generalise a theoretical framework elaborated by ourselves to identify 'quantum mechanical statistics' in large-size texts. Then, we show that, in all analysed texts, words distribute according to 'Bose--Einstein statistics' and show significant deviations from 'Maxwell--Boltzmann statistics'. Next, we introduce an effect of 'word randomization' which instead indicates that the difference between the two statistical models is not as pronounced as in the original cases. These results confirm the empirical patterns obtained in texts of English language and strongly indicate that identical words tend to 'clump together' as a consequence of their meaning, which can be explained as an effect of 'quantum entanglement' produced through a phenomenon of 'contextual updating'. More, word randomization can be seen as the linguistic-conceptual equivalent of an increase of temperature which destroys 'coherence' and makes classical statistics prevail over quantum statistics. Some insights into the origin of quantum statistics in physics are finally provided.
comment: 21 pages, 6 figures
☆ Rethinking Emotion Annotations in the Era of Large Language Models
Modern affective computing systems rely heavily on datasets with human-annotated emotion labels, for training and evaluation. However, human annotations are expensive to obtain, sensitive to study design, and difficult to quality control, because of the subjective nature of emotions. Meanwhile, Large Language Models (LLMs) have shown remarkable performance on many Natural Language Understanding tasks, emerging as a promising tool for text annotation. In this work, we analyze the complexities of emotion annotation in the context of LLMs, focusing on GPT-4 as a leading model. In our experiments, GPT-4 achieves high ratings in a human evaluation study, painting a more positive picture than previous work, in which human labels served as the only ground truth. On the other hand, we observe differences between human and GPT-4 emotion perception, underscoring the importance of human input in annotation studies. To harness GPT-4's strength while preserving human perspective, we explore two ways of integrating GPT-4 into emotion annotation pipelines, showing its potential to flag low-quality labels, reduce the workload of human annotators, and improve downstream model learning performance and efficiency. Together, our findings highlight opportunities for new emotion labeling practices and suggest the use of LLMs as a promising tool to aid human annotation.
♻ ☆ Beyond Retrieval: Generating Narratives in Conversational Recommender Systems
The recent advances in Large Language Model's generation and reasoning capabilities present an opportunity to develop truly conversational recommendation systems. However, effectively integrating recommender system knowledge into LLMs for natural language generation which is tailored towards recommendation tasks remains a challenge. This paper addresses this challenge by making two key contributions. First, we introduce a new dataset (REGEN) for natural language generation tasks in conversational recommendations. REGEN (Reviews Enhanced with GEnerative Narratives) extends the Amazon Product Reviews dataset with rich user narratives, including personalized explanations of product preferences, product endorsements for recommended items, and summaries of user purchase history. REGEN is made publicly available to facilitate further research. Furthermore, we establish benchmarks using well-known generative metrics, and perform an automated evaluation of the new dataset using a rater LLM. Second, the paper introduces a fusion architecture (CF model with an LLM) which serves as a baseline for REGEN. And to the best of our knowledge, represents the first attempt to analyze the capabilities of LLMs in understanding recommender signals and generating rich narratives. We demonstrate that LLMs can effectively learn from simple fusion architectures utilizing interaction-based CF embeddings, and this can be further enhanced using the metadata and personalization data associated with items. Our experiments show that combining CF and content embeddings leads to improvements of 4-12% in key language metrics compared to using either type of embedding individually. We also provide an analysis to interpret how CF and content embeddings contribute to this new generative task.
♻ ☆ Toward Self-Improvement of LLMs via Imagination, Searching, and Criticizing NeurIPS 2024
Despite the impressive capabilities of Large Language Models (LLMs) on various tasks, they still struggle with scenarios that involves complex reasoning and planning. Recent work proposed advanced prompting techniques and the necessity of fine-tuning with high-quality data to augment LLMs' reasoning abilities. However, these approaches are inherently constrained by data availability and quality. In light of this, self-correction and self-learning emerge as viable solutions, employing strategies that allow LLMs to refine their outputs and learn from self-assessed rewards. Yet, the efficacy of LLMs in self-refining its response, particularly in complex reasoning and planning task, remains dubious. In this paper, we introduce AlphaLLM for the self-improvements of LLMs, which integrates Monte Carlo Tree Search (MCTS) with LLMs to establish a self-improving loop, thereby enhancing the capabilities of LLMs without additional annotations. Drawing inspiration from the success of AlphaGo, AlphaLLM addresses the unique challenges of combining MCTS with LLM for self-improvement, including data scarcity, the vastness search spaces of language tasks, and the subjective nature of feedback in language tasks. AlphaLLM is comprised of prompt synthesis component, an efficient MCTS approach tailored for language tasks, and a trio of critic models for precise feedback. Our experimental results in mathematical reasoning tasks demonstrate that AlphaLLM significantly enhances the performance of LLMs without additional annotations, showing the potential for self-improvement in LLMs.
comment: NeurIPS 2024
♻ ☆ Scalable Influence and Fact Tracing for Large Language Model Pretraining
Training data attribution (TDA) methods aim to attribute model outputs back to specific training examples, and the application of these methods to large language model (LLM) outputs could significantly advance model transparency and data curation. However, it has been challenging to date to apply these methods to the full scale of LLM pretraining. In this paper, we refine existing gradient-based methods to work effectively at scale, allowing us to retrieve influential examples for an 8B-parameter language model from a pretraining corpus of over 160B tokens with no need for subsampling or pre-filtering. Our method combines several techniques, including optimizer state correction, a task-specific Hessian approximation, and normalized encodings, which we find to be critical for performance at scale. In quantitative evaluations on a fact tracing task, our method performs best at identifying examples that influence model predictions, but classical, model-agnostic retrieval methods such as BM25 still perform better at finding passages which explicitly contain relevant facts. These results demonstrate a misalignment between factual *attribution* and causal *influence*. With increasing model size and training tokens, we find that influence more closely aligns with factual attribution. Finally, we examine different types of examples identified as influential by our method, finding that while many directly entail a particular fact, others support the same output by reinforcing priors on relation types, common entities, and names. We release our prompt set and model outputs, along with a web-based visualization tool to explore influential examples for factual predictions, commonsense reasoning, arithmetic, and open-ended generation for an 8B-parameter LLM.
♻ ☆ An Entailment Tree Generation Approach for Multimodal Multi-Hop Question Answering with Mixture-of-Experts and Iterative Feedback Mechanism
With the rise of large-scale language models (LLMs), it is currently popular and effective to convert multimodal information into text descriptions for multimodal multi-hop question answering. However, we argue that the current methods of multi-modal multi-hop question answering still mainly face two challenges: 1) The retrieved evidence containing a large amount of redundant information, inevitably leads to a significant drop in performance due to irrelevant information misleading the prediction. 2) The reasoning process without interpretable reasoning steps makes the model difficult to discover the logical errors for handling complex questions. To solve these problems, we propose a unified LLMs-based approach but without heavily relying on them due to the LLM's potential errors, and innovatively treat multimodal multi-hop question answering as a joint entailment tree generation and question answering problem. Specifically, we design a multi-task learning framework with a focus on facilitating common knowledge sharing across interpretability and prediction tasks while preventing task-specific errors from interfering with each other via mixture of experts. Afterward, we design an iterative feedback mechanism to further enhance both tasks by feeding back the results of the joint training to the LLM for regenerating entailment trees, aiming to iteratively refine the potential answer. Notably, our method has won the first place in the official leaderboard of WebQA (since April 10, 2024), and achieves competitive results on MultimodalQA.
comment: Erratum: We identified an error in the calculation of the F1 score in table 4 reported in a previous version of this work. The performance of the new result is better than the previous one. The corrected values are included in this updated version of the paper. These changes do not alter the primary conclusions of our research
♻ ☆ Which LLMs are Difficult to Detect? A Detailed Analysis of Potential Factors Contributing to Difficulties in LLM Text Detection NeurIPS 2024
As LLMs increase in accessibility, LLM-generated texts have proliferated across several fields, such as scientific, academic, and creative writing. However, LLMs are not created equally; they may have different architectures and training datasets. Thus, some LLMs may be more challenging to detect than others. Using two datasets spanning four total writing domains, we train AI-generated (AIG) text classifiers using the LibAUC library - a deep learning library for training classifiers with imbalanced datasets. Our results in the Deepfake Text dataset show that AIG-text detection varies across domains, with scientific writing being relatively challenging. In the Rewritten Ivy Panda (RIP) dataset focusing on student essays, we find that the OpenAI family of LLMs was substantially difficult for our classifiers to distinguish from human texts. Additionally, we explore possible factors that could explain the difficulties in detecting OpenAI-generated texts.
comment: Accepted at NeurIPS 2024 - Safe Generative AI Workshop; Camera-ready version
♻ ☆ GenderAlign: An Alignment Dataset for Mitigating Gender Bias in Large Language Models
Large Language Models (LLMs) are prone to generating content that exhibits gender biases, raising significant ethical concerns. Alignment, the process of fine-tuning LLMs to better align with desired behaviors, is recognized as an effective approach to mitigate gender biases. Although proprietary LLMs have made significant strides in mitigating gender bias, their alignment datasets are not publicly available. The commonly used and publicly available alignment dataset, HH-RLHF, still exhibits gender bias to some extent. There is a lack of publicly available alignment datasets specifically designed to address gender bias. Hence, we developed a new dataset named GenderAlign, aiming at mitigating a comprehensive set of gender biases in LLMs. This dataset comprises 8k single-turn dialogues, each paired with a "chosen" and a "rejected" response. Compared to the "rejected" responses, the "chosen" responses demonstrate lower levels of gender bias and higher quality. Furthermore, we categorized the gender biases in the "rejected" responses of GenderAlign into 4 principal categories. The experimental results show the effectiveness of GenderAlign in reducing gender bias in LLMs.
♻ ☆ CPsyExam: A Chinese Benchmark for Evaluating Psychology using Examinations COLING 2025
In this paper, we introduce a novel psychological benchmark, CPsyExam, constructed from questions sourced from Chinese language examinations. CPsyExam is designed to prioritize psychological knowledge and case analysis separately, recognizing the significance of applying psychological knowledge to real-world scenarios. From the pool of 22k questions, we utilize 4k to create the benchmark that offers balanced coverage of subjects and incorporates a diverse range of case analysis techniques.Furthermore, we evaluate a range of existing large language models~(LLMs), spanning from open-sourced to API-based models. Our experiments and analysis demonstrate that CPsyExam serves as an effective benchmark for enhancing the understanding of psychology within LLMs and enables the comparison of LLMs across various granularities.
comment: To appear in COLING 2025
♻ ☆ Guiding Vision-Language Model Selection for Visual Question-Answering Across Tasks, Domains, and Knowledge Types
Visual Question-Answering (VQA) has become key to user experience, particularly after improved generalization capabilities of Vision-Language Models (VLMs). But evaluating VLMs for an application requirement using a standardized framework in practical settings is still challenging. This paper aims to solve that using an end-to-end framework. We present VQA360 - a novel dataset derived from established VQA benchmarks, annotated with task types, application domains, and knowledge types, for a comprehensive evaluation. We also introduce GoEval, a multimodal evaluation metric developed using GPT-4o, achieving a correlation factor of 56.71% with human judgments. Our experiments with state-of-the-art VLMs reveal that no single model excels universally, thus, making a right choice a key design decision. Proprietary models such as Gemini-1.5-Pro and GPT-4o-mini generally outperform others, but open-source models like InternVL-2-8B and CogVLM-2-Llama-3-19B also demonstrate competitive strengths, while providing additional advantages. Our framework can also be extended to other tasks.
comment: 8 pages + references + 6 pages of Appendix
♻ ☆ Let Guidelines Guide You: A Prescriptive Guideline-Centered Data Annotation Methodology
We introduce the Guideline-Centered Annotation Methodology (GCAM), a novel data annotation methodology designed to report the annotation guidelines associated with each data sample. Our approach addresses three key limitations of the standard prescriptive annotation methodology by reducing the information loss during annotation and ensuring adherence to guidelines. Furthermore, GCAM enables the efficient reuse of annotated data across multiple tasks. We evaluate GCAM in two ways: (i) through a human annotation study and (ii) an experimental evaluation with several machine learning models. Our results highlight the advantages of GCAM from multiple perspectives, demonstrating its potential to improve annotation quality and error analysis.
♻ ☆ AutoDetect: Towards a Unified Framework for Automated Weakness Detection in Large Language Models EMNLP 2024
Although Large Language Models (LLMs) are becoming increasingly powerful, they still exhibit significant but subtle weaknesses, such as mistakes in instruction-following or coding tasks. As these unexpected errors could lead to severe consequences in practical deployments, it is crucial to investigate the limitations within LLMs systematically. Traditional benchmarking approaches cannot thoroughly pinpoint specific model deficiencies, while manual inspections are costly and not scalable. In this paper, we introduce a unified framework, AutoDetect, to automatically expose weaknesses in LLMs across various tasks. Inspired by the educational assessment process that measures students' learning outcomes, AutoDetect consists of three LLM-powered agents: Examiner, Questioner, and Assessor. The collaboration among these three agents is designed to realize comprehensive and in-depth weakness identification. Our framework demonstrates significant success in uncovering flaws, with an identification success rate exceeding 30% in prominent models such as ChatGPT and Claude. More importantly, these identified weaknesses can guide specific model improvements, proving more effective than untargeted data augmentation methods like Self-Instruct. Our approach has led to substantial enhancements in popular LLMs, including the Llama series and Mistral-7b, boosting their performance by over 10% across several benchmarks. Code and data are publicly available at https://github.com/thu-coai/AutoDetect.
comment: EMNLP 2024 findings
♻ ☆ Enhancing Vision-Language Model Pre-training with Image-text Pair Pruning Based on Word Frequency
We propose Word-Frequency-based Image-Text Pair Pruning (WFPP), a novel data pruning method that improves the efficiency of VLMs. Unlike MetaCLIP, our method does not need metadata for pruning, but selects text-image pairs to prune based on the content of the text. Specifically, WFPP prunes text-image pairs containing high-frequency words across the entire training dataset. The effect of WFPP is to reduce the dominance of frequent words. The result a better balanced word-frequency distribution in the dataset, which is known to improve the training of word embedding models. After pre-training on the pruned subset, we fine-tuned the model on the entire dataset for one additional epoch to achieve better performance. Our experiments demonstrate that applying WFPP when training a CLIP model improves performance on a wide range of downstream tasks. WFPP also provides the advantage of speeding up pre-training by using fewer samples. Additionally, we analyze the training data before and after pruning to visualize how WFPP changes the balance of word frequencies. We hope our work encourages researchers to consider the distribution of words in the training data when pre-training VLMs, not limited to CLIP.
♻ ☆ Automatic register identification for the open web using multilingual deep learning
This article investigates how well deep learning models can identify web registers -- text varieties such as news reports and discussion forums -- across 16 languages. We introduce the Multilingual CORE corpora, which contain 72,504 documents annotated with a hierarchical taxonomy of 25 registers designed to cover the entire open web. Our multilingual models achieve state-of-the-art results (79% F1 score) using multi-label classification. This performance matches or exceeds previous studies that used simpler classification schemes, showing that models can perform well even with a complex register scheme at a massively multilingual scale. However, we observe a consistent performance ceiling around 77-80% F1 score across all models and configurations. When we remove documents with uncertain labels through data pruning, performance increases to over 90% F1, suggesting that this ceiling stems from inherent ambiguity in web registers rather than model limitations. Analysis of hybrid documents -- texts combining multiple registers -- reveals that the main challenge is not in classifying hybrids themselves, but in distinguishing between hybrid and non-hybrid documents. Multilingual models consistently outperform monolingual ones, particularly helping languages with limited training data. While zero-shot performance drops by an average of 7% on unseen languages, this decrease varies substantially between languages (from 3% to 20%), indicating that while registers share many features across languages, they also maintain language-specific characteristics.
♻ ☆ VP-MEL: Visual Prompts Guided Multimodal Entity Linking
Multimodal Entity Linking (MEL) is extensively utilized in the domains of information retrieval. However, existing MEL methods typically utilize mention words as mentions for retrieval. This results in a significant dependence of MEL on mention words, thereby constraining its capacity to effectively leverage information from both images and text. In situations where mention words are absent, MEL methods struggle to leverage image-text pairs for entity linking. To solve these issues, we introduce a Visual Prompts guided Multimodal Entity Linking (VP-MEL) task. VP-MEL directly marks specific regions within the image. These markers are referred to as visual prompts in VP-MEL. Without mention words, VP-MEL aims to utilize marked image-text pairs to align visual prompts with specific entities in the knowledge bases. A new dataset for the VP-MEL task, VPWiki, is proposed in this paper. Moreover, we propose a framework named FBMEL, which enhances the significance of visual prompts and fully leverages the information in image-text pairs. Experimental results on the VPWiki dataset demonstrate that FBMEL outperforms baseline methods across multiple benchmarks for the VP-MEL task.
♻ ☆ MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
Autonomous agents powered by large language models (LLMs) show promising potential in assistive tasks across various domains, including mobile device control. As these agents interact directly with personal information and device settings, ensuring their safe and reliable behavior is crucial to prevent undesirable outcomes. However, no benchmark exists for standardized evaluation of the safety of mobile device-control agents. In this work, we introduce MobileSafetyBench, a benchmark designed to evaluate the safety of device-control agents within a realistic mobile environment based on Android emulators. We develop a diverse set of tasks involving interactions with various mobile applications, including messaging and banking applications, challenging agents with managing risks encompassing misuse and negative side effects. These tasks include tests to evaluate the safety of agents in daily scenarios as well as their robustness against indirect prompt injection attacks. Our experiments demonstrate that baseline agents, based on state-of-the-art LLMs, often fail to effectively prevent harm while performing the tasks. To mitigate these safety concerns, we propose a prompting method that encourages agents to prioritize safety considerations. While this method shows promise in promoting safer behaviors, there is still considerable room for improvement to fully earn user trust. This highlights the urgent need for continued research to develop more robust safety mechanisms in mobile environments. We open-source our benchmark at: https://mobilesafetybench.github.io/.
♻ ☆ Using Large Language Models for Expert Prior Elicitation in Predictive Modelling
Large language models (LLMs), trained on diverse data effectively acquire a breadth of information across various domains. However, their computational complexity, cost, and lack of transparency hinder their direct application for specialised tasks. In fields such as clinical research, acquiring expert annotations or prior knowledge about predictive models is often costly and time-consuming. This study proposes the use of LLMs to elicit expert prior distributions for predictive models. This approach also provides an alternative to in-context learning, where language models are tasked with making predictions directly. In this work, we compare LLM-elicited and uninformative priors, evaluate whether LLMs truthfully generate parameter distributions, and propose a model selection strategy for in-context learning and prior elicitation. Our findings show that LLM-elicited prior parameter distributions significantly reduce predictive error compared to uninformative priors in low-data settings. Applied to clinical problems, this translates to fewer required biological samples, lowering cost and resources. Prior elicitation also consistently outperforms and proves more reliable than in-context learning at a lower cost, making it a preferred alternative in our setting. We demonstrate the utility of this method across various use cases, including clinical applications. For infection prediction, using LLM-elicited priors reduced the number of required labels to achieve the same accuracy as an uninformative prior by 55%, 200 days earlier in the study.
♻ ☆ Is Self-knowledge and Action Consistent or Not: Investigating Large Language Model's Personality ICML 2024
In this study, we delve into the validity of conventional personality questionnaires in capturing the human-like personality traits of Large Language Models (LLMs). Our objective is to assess the congruence between the personality traits LLMs claim to possess and their demonstrated tendencies in real-world scenarios. By conducting an extensive examination of LLM outputs against observed human response patterns, we aim to understand the disjunction between self-knowledge and action in LLMs.
comment: ICML 2024, Large Language Models and Cognition
♻ ☆ From Measurement Instruments to Data: Leveraging Theory-Driven Synthetic Training Data for Classifying Social Constructs
Computational text classification is a challenging task, especially for multi-dimensional social constructs. Recently, there has been increasing discussion that synthetic training data could enhance classification by offering examples of how these constructs are represented in texts. In this paper, we systematically examine the potential of theory-driven synthetic training data for improving the measurement of social constructs. In particular, we explore how researchers can transfer established knowledge from measurement instruments in the social sciences, such as survey scales or annotation codebooks, into theory-driven generation of synthetic data. Using two studies on measuring sexism and political topics, we assess the added value of synthetic training data for fine-tuning text classification models. Although the results of the sexism study were less promising, our findings demonstrate that synthetic data can be highly effective in reducing the need for labeled data in political topic classification. With only a minimal drop in performance, synthetic data allows for substituting large amounts of labeled data. Furthermore, theory-driven synthetic data performed markedly better than data generated without conceptual information in mind.
♻ ☆ Benchmarking Chinese Commonsense Reasoning of LLMs: From Chinese-Specifics to Reasoning-Memorization Correlations
We introduce CHARM, the first benchmark for comprehensively and in-depth evaluating the commonsense reasoning ability of large language models (LLMs) in Chinese, which covers both globally known and Chinese-specific commonsense. We evaluated 7 English and 12 Chinese-oriented LLMs on CHARM, employing 5 representative prompt strategies for improving LLMs' reasoning ability, such as Chain-of-Thought. Our findings indicate that the LLM's language orientation and the task's domain influence the effectiveness of the prompt strategy, which enriches previous research findings. We built closely-interconnected reasoning and memorization tasks, and found that some LLMs struggle with memorizing Chinese commonsense, affecting their reasoning ability, while others show differences in reasoning despite similar memorization performance. We also evaluated the LLMs' memorization-independent reasoning abilities and analyzed the typical errors. Our study precisely identified the LLMs' strengths and weaknesses, providing the clear direction for optimization. It can also serve as a reference for studies in other fields. We will release CHARM at https://github.com/opendatalab/CHARM .
comment: Equal contribution: Jiaxing Sun, Weiquan Huang, Jiang Wu; Corresponding author: Conghui He
♻ ☆ Mixture of Hidden-Dimensions Transformer
Transformer models encounter challenges in scaling hidden dimensions efficiently, as uniformly increasing them inflates computational and memory costs while failing to emphasize the most relevant features for each token. For further understanding, we study hidden dimension sparsity and observe that trained Transformers utilize only a small fraction of token dimensions, revealing an "activation flow" pattern. Notably, there are shared sub-dimensions with sustained activation across multiple consecutive tokens and specialized sub-dimensions uniquely activated for each token. To better model token-relevant sub-dimensions, we propose MoHD (Mixture of Hidden Dimensions), a sparse conditional activation architecture. Particularly, MoHD employs shared sub-dimensions for common token features and a routing mechanism to dynamically activate specialized sub-dimensions. To mitigate potential information loss from sparsity, we design activation scaling and group fusion mechanisms to preserve activation flow. In this way, MoHD expands hidden dimensions with negligible increases in computation or parameters, efficient training and inference while maintaining performance. Evaluations across 10 NLP tasks show that MoHD surpasses Vanilla Transformers in parameter efficiency and task performance. It achieves 1.7% higher performance with 50% fewer activation parameters and 3.7% higher performance with a 3x parameter expansion at constant activation cost. MOHD offers a new perspective for scaling the model, showcasing the potential of hidden dimension sparsity to boost efficiency
comment: 16 pages, 10 figures, 5 tables
♻ ☆ RST-LoRA: A Discourse-Aware Low-Rank Adaptation for Long Document Abstractive Summarization NAACL 2024
For long document summarization, discourse structure is important to discern the key content of the text and the differences in importance level between sentences. Unfortunately, the integration of rhetorical structure theory (RST) into parameter-efficient fine-tuning strategies for long document summarization remains unexplored. Therefore, this paper introduces RST-LoRA and proposes four RST-aware variants to explicitly incorporate RST into the LoRA model. Our empirical evaluation demonstrates that incorporating the type and uncertainty of rhetorical relations can complementarily enhance the performance of LoRA in summarization tasks. Furthermore, the best-performing variant we introduced outperforms the vanilla LoRA and full-parameter fine-tuning models, as confirmed by multiple automatic and human evaluations, and even surpasses previous state-of-the-art methods.
comment: NAACL 2024 Main & Long Conference Paper (Oral Presentation)
♻ ☆ SciNews: From Scholarly Complexities to Public Narratives -- A Dataset for Scientific News Report Generation LREC
Scientific news reports serve as a bridge, adeptly translating complex research articles into reports that resonate with the broader public. The automated generation of such narratives enhances the accessibility of scholarly insights. In this paper, we present a new corpus to facilitate this paradigm development. Our corpus comprises a parallel compilation of academic publications and their corresponding scientific news reports across nine disciplines. To demonstrate the utility and reliability of our dataset, we conduct an extensive analysis, highlighting the divergences in readability and brevity between scientific news narratives and academic manuscripts. We benchmark our dataset employing state-of-the-art text generation models. The evaluation process involves both automatic and human evaluation, which lays the groundwork for future explorations into the automated generation of scientific news reports. The dataset and code related to this work are available at https://dongqi.me/projects/SciNews.
comment: LREC-COLING 2024 Main Conference Paper
♻ ☆ ChatGPT vs Human-authored Text: Insights into Controllable Text Summarization and Sentence Style Transfer ACL
Large-scale language models, like ChatGPT, have garnered significant media attention and stunned the public with their remarkable capacity for generating coherent text from short natural language prompts. In this paper, we aim to conduct a systematic inspection of ChatGPT's performance in two controllable generation tasks, with respect to ChatGPT's ability to adapt its output to different target audiences (expert vs. layman) and writing styles (formal vs. informal). Additionally, we evaluate the faithfulness of the generated text, and compare the model's performance with human-authored texts. Our findings indicate that the stylistic variations produced by humans are considerably larger than those demonstrated by ChatGPT, and the generated texts diverge from human samples in several characteristics, such as the distribution of word types. Moreover, we observe that ChatGPT sometimes incorporates factual errors or hallucinations when adapting the text to suit a specific style.
comment: ACL-SRW 2023
♻ ☆ Incorporating Distributions of Discourse Structure for Long Document Abstractive Summarization ACL 2023
For text summarization, the role of discourse structure is pivotal in discerning the core content of a text. Regrettably, prior studies on incorporating Rhetorical Structure Theory (RST) into transformer-based summarization models only consider the nuclearity annotation, thereby overlooking the variety of discourse relation types. This paper introduces the 'RSTformer', a novel summarization model that comprehensively incorporates both the types and uncertainty of rhetorical relations. Our RST-attention mechanism, rooted in document-level rhetorical structure, is an extension of the recently devised Longformer framework. Through rigorous evaluation, the model proposed herein exhibits significant superiority over state-of-the-art models, as evidenced by its notable performance on several automatic metrics and human evaluation.
comment: ACL 2023 (Main conference)
♻ ☆ Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision NeurIPS 2024
Current AI alignment methodologies rely on human-provided demonstrations or judgments, and the learned capabilities of AI systems would be upper-bounded by human capabilities as a result. This raises a challenging research question: How can we keep improving the systems when their capabilities have surpassed the levels of humans? This paper answers this question in the context of tackling hard reasoning tasks (e.g., level 4-5 MATH problems) via learning from human annotations on easier tasks (e.g., level 1-3 MATH problems), which we term as easy-to-hard generalization. Our key insight is that an evaluator (reward model) trained on supervisions for easier tasks can be effectively used for scoring candidate solutions of harder tasks and hence facilitating easy-to-hard generalization over different levels of tasks. Based on this insight, we propose a novel approach to scalable alignment, which firstly trains the (process-supervised) reward models on easy problems (e.g., level 1-3), and then uses them to evaluate the performance of policy models on hard problems. We show that such easy-to-hard generalization from evaluators can enable easy-to-hard generalizations in generators either through re-ranking or reinforcement learning (RL). Notably, our process-supervised 7b RL model and 34b model (reranking@1024) achieves an accuracy of 34.0% and 52.5% on MATH500, respectively, despite only using human supervision on easy problems. Our approach suggests a promising path toward AI systems that advance beyond the frontier of human supervision.
comment: Accepted at NeurIPS 2024
♻ ☆ ProcessBench: Identifying Process Errors in Mathematical Reasoning
As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.
♻ ☆ Reprint: a randomized extrapolation based on principal components for data augmentation
Data scarcity and data imbalance have attracted a lot of attention in many fields. Data augmentation, explored as an effective approach to tackle them, can improve the robustness and efficiency of classification models by generating new samples. This paper presents REPRINT, a simple and effective hidden-space data augmentation method for imbalanced data classification. Given hidden-space representations of samples in each class, REPRINT extrapolates, in a randomized fashion, augmented examples for target class by using subspaces spanned by principal components to summarize distribution structure of both source and target class. Consequently, the examples generated would diversify the target while maintaining the original geometry of target distribution. Besides, this method involves a label refinement component which allows to synthesize new soft labels for augmented examples. Compared with different NLP data augmentation approaches under a range of data imbalanced scenarios on four text classification benchmark, REPRINT shows prominent improvements. Moreover, through comprehensive ablation studies, we show that label refinement is better than label-preserving for augmented examples, and that our method suggests stable and consistent improvements in terms of suitable choices of principal components. Moreover, REPRINT is appealing for its easy-to-use since it contains only one hyperparameter determining the dimension of subspace and requires low computational resource.
♻ ☆ LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods
The rapid advancement of Large Language Models (LLMs) has driven their expanding application across various fields. One of the most promising applications is their role as evaluators based on natural language responses, referred to as ''LLMs-as-judges''. This framework has attracted growing attention from both academia and industry due to their excellent effectiveness, ability to generalize across tasks, and interpretability in the form of natural language. This paper presents a comprehensive survey of the LLMs-as-judges paradigm from five key perspectives: Functionality, Methodology, Applications, Meta-evaluation, and Limitations. We begin by providing a systematic definition of LLMs-as-Judges and introduce their functionality (Why use LLM judges?). Then we address methodology to construct an evaluation system with LLMs (How to use LLM judges?). Additionally, we investigate the potential domains for their application (Where to use LLM judges?) and discuss methods for evaluating them in various contexts (How to evaluate LLM judges?). Finally, we provide a detailed analysis of the limitations of LLM judges and discuss potential future directions. Through a structured and comprehensive analysis, we aim aims to provide insights on the development and application of LLMs-as-judges in both research and practice. We will continue to maintain the relevant resource list at https://github.com/CSHaitao/Awesome-LLMs-as-Judges.
comment: 60 pages, comprehensive and continuously updated
♻ ☆ From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge
Assessment and evaluation have long been critical challenges in artificial intelligence (AI) and natural language processing (NLP). However, traditional methods, whether matching-based or embedding-based, often fall short of judging subtle attributes and delivering satisfactory results. Recent advancements in Large Language Models (LLMs) inspire the "LLM-as-a-judge" paradigm, where LLMs are leveraged to perform scoring, ranking, or selection across various tasks and applications. This paper provides a comprehensive survey of LLM-based judgment and assessment, offering an in-depth overview to advance this emerging field. We begin by giving detailed definitions from both input and output perspectives. Then we introduce a comprehensive taxonomy to explore LLM-as-a-judge from three dimensions: what to judge, how to judge and where to judge. Finally, we compile benchmarks for evaluating LLM-as-a-judge and highlight key challenges and promising directions, aiming to provide valuable insights and inspire future research in this promising research area. Paper list and more resources about LLM-as-a-judge can be found at \url{https://github.com/llm-as-a-judge/Awesome-LLM-as-a-judge} and \url{https://llm-as-a-judge.github.io}.
comment: v2: add missing citations; 32 pages, 5 figures
♻ ☆ Data Quality Enhancement on the Basis of Diversity with Large Language Models for Text Classification: Uncovered, Difficult, and Noisy COLING 2025
In recent years, the use of large language models (LLMs) for text classification has attracted widespread attention. Despite this, the classification accuracy of LLMs has not yet universally surpassed that of smaller models. LLMs can enhance their performance in text classification through fine-tuning. However, existing data quality research based on LLMs is challenging to apply directly to solve text classification problems. To further improve the performance of LLMs in classification tasks, this paper proposes a data quality enhancement (DQE) method for text classification based on LLMs. This method starts by using a greedy algorithm to select data, dividing the dataset into sampled and unsampled subsets, and then performing fine-tuning of the LLMs using the sampled data. Subsequently, this model is used to predict the outcomes for the unsampled data, categorizing incorrectly predicted data into uncovered, difficult, and noisy data. Experimental results demonstrate that our method effectively enhances the performance of LLMs in text classification tasks and significantly improves training efficiency, saving nearly half of the training time. Our method has achieved state-of-the-art performance in several open-source classification tasks.
comment: Accepted by COLING 2025(main, long paper)
♻ ☆ Acquiring Bidirectionality via Large and Small Language Models COLING2025
Using token representation from bidirectional language models (LMs) such as BERT is still a widely used approach for token-classification tasks. Even though there exist much larger unidirectional LMs such as Llama-2, they are rarely used to replace the token representation of bidirectional LMs. In this work, we hypothesize that their lack of bidirectionality is keeping them behind. To that end, we propose to newly train a small backward LM and concatenate its representations to those of existing LM for downstream tasks. Through experiments in named entity recognition, we demonstrate that introducing backward model improves the benchmark performance more than 10 points. Furthermore, we show that the proposed method is especially effective for rare domains and in few-shot learning settings.
comment: Accepted by COLING2025
♻ ☆ Sortformer: Seamless Integration of Speaker Diarization and ASR by Bridging Timestamps and Tokens
We propose Sortformer, a novel neural model for speaker diarization, trained with unconventional objectives compared to existing end-to-end diarization models. The permutation problem in speaker diarization has long been regarded as a critical challenge. Most prior end-to-end diarization systems employ permutation invariant loss (PIL), which optimizes for the permutation that yields the lowest error. In contrast, we introduce Sort Loss, which enables a diarization model to autonomously resolve permutation, with or without PIL. We demonstrate that combining Sort Loss and PIL achieves performance competitive with state-of-the-art end-to-end diarization models trained exclusively with PIL. Crucially, we present a streamlined multispeaker ASR architecture that leverages Sortformer as a speaker supervision model, embedding speaker label estimation within the ASR encoder state using a sinusoidal kernel function. This approach resolves the speaker permutation problem through sorted objectives, effectively bridging speaker-label timestamps and speaker tokens. In our experiments, we show that the proposed multispeaker ASR architecture, enhanced with speaker supervision, improves performance via adapter techniques. Code and trained models will be made publicly available via the NVIDIA NeMo framework.
♻ ☆ From Text to Multimodality: Exploring the Evolution and Impact of Large Language Models in Medical Practice
Large Language Models (LLMs) have rapidly evolved from text-based systems to multimodal platforms, significantly impacting various sectors including healthcare. This comprehensive review explores the progression of LLMs to Multimodal Large Language Models (MLLMs) and their growing influence in medical practice. We examine the current landscape of MLLMs in healthcare, analyzing their applications across clinical decision support, medical imaging, patient engagement, and research. The review highlights the unique capabilities of MLLMs in integrating diverse data types, such as text, images, and audio, to provide more comprehensive insights into patient health. We also address the challenges facing MLLM implementation, including data limitations, technical hurdles, and ethical considerations. By identifying key research gaps, this paper aims to guide future investigations in areas such as dataset development, modality alignment methods, and the establishment of ethical guidelines. As MLLMs continue to shape the future of healthcare, understanding their potential and limitations is crucial for their responsible and effective integration into medical practice.
comment: 12 pages, 1 figure
♻ ☆ A Dataset and Benchmark for Hospital Course Summarization with Adapted Large Language Models
Brief hospital course (BHC) summaries are clinical documents that summarize a patient's hospital stay. While large language models (LLMs) depict remarkable capabilities in automating real-world tasks, their capabilities for healthcare applications such as synthesizing BHCs from clinical notes have not been shown. We introduce a novel pre-processed dataset, the MIMIC-IV-BHC, encapsulating clinical note and brief hospital course (BHC) pairs to adapt LLMs for BHC synthesis. Furthermore, we introduce a benchmark of the summarization performance of two general-purpose LLMs and three healthcare-adapted LLMs. Using clinical notes as input, we apply prompting-based (using in-context learning) and fine-tuning-based adaptation strategies to three open-source LLMs (Clinical-T5-Large, Llama2-13B, FLAN-UL2) and two proprietary LLMs (GPT-3.5, GPT-4). We evaluate these LLMs across multiple context-length inputs using natural language similarity metrics. We further conduct a clinical study with five clinicians, comparing clinician-written and LLM-generated BHCs across 30 samples, focusing on their potential to enhance clinical decision-making through improved summary quality. We observe that the Llama2-13B fine-tuned LLM outperforms other domain-adapted models given quantitative evaluation metrics of BLEU and BERT-Score. GPT-4 with in-context learning shows more robustness to increasing context lengths of clinical note inputs than fine-tuned Llama2-13B. Despite comparable quantitative metrics, the reader study depicts a significant preference for summaries generated by GPT-4 with in-context learning compared to both Llama2-13B fine-tuned summaries and the original summaries, highlighting the need for qualitative clinical evaluation.
♻ ☆ WaterSeeker: Pioneering Efficient Detection of Watermarked Segments in Large Documents
Watermarking algorithms for large language models (LLMs) have attained high accuracy in detecting LLM-generated text. However, existing methods primarily focus on distinguishing fully watermarked text from non-watermarked text, overlooking real-world scenarios where LLMs generate only small sections within large documents. In this scenario, balancing time complexity and detection performance poses significant challenges. This paper presents WaterSeeker, a novel approach to efficiently detect and locate watermarked segments amid extensive natural text. It first applies an efficient anomaly extraction method to preliminarily locate suspicious watermarked regions. Following this, it conducts a local traversal and performs full-text detection for more precise verification. Theoretical analysis and experimental results demonstrate that WaterSeeker achieves a superior balance between detection accuracy and computational efficiency. Moreover, its localization capability lays the foundation for building interpretable AI detection systems. Our code is available at https://github.com/THU-BPM/WaterSeeker.
comment: 17 pages, 6 figures, 7 tables
♻ ☆ REBEL: Reinforcement Learning via Regressing Relative Rewards
While originally developed for continuous control problems, Proximal Policy Optimization (PPO) has emerged as the work-horse of a variety of reinforcement learning (RL) applications, including the fine-tuning of generative models. Unfortunately, PPO requires multiple heuristics to enable stable convergence (e.g. value networks, clipping), and is notorious for its sensitivity to the precise implementation of these components. In response, we take a step back and ask what a minimalist RL algorithm for the era of generative models would look like. We propose REBEL, an algorithm that cleanly reduces the problem of policy optimization to regressing the relative reward between two completions to a prompt in terms of the policy, enabling strikingly lightweight implementation. In theory, we prove that fundamental RL algorithms like Natural Policy Gradient can be seen as variants of REBEL, which allows us to match the strongest known theoretical guarantees in terms of convergence and sample complexity in the RL literature. REBEL can also cleanly incorporate offline data and be extended to handle the intransitive preferences we frequently see in practice. Empirically, we find that REBEL provides a unified approach to language modeling and image generation with stronger or similar performance as PPO and DPO, all while being simpler to implement and more computationally efficient than PPO. When fine-tuning Llama-3-8B-Instruct, REBEL achieves strong performance in AlpacaEval 2.0, MT-Bench, and Open LLM Leaderboard.
comment: New experimental results on general chat
♻ ☆ CFSP: An Efficient Structured Pruning Framework for LLMs with Coarse-to-Fine Activation Information COLING2025
The colossal parameters and computational overhead of Large Language Models (LLMs) challenge their real-world applications. Network pruning, which targets unstructured or structured sparsity by removing redundant parameters, has recently been explored for LLM acceleration. Existing LLM pruning works focus on unstructured pruning, which typically requires special hardware support for a practical speed-up. In contrast, structured pruning can reduce latency on general devices. However, it remains a challenge to perform structured pruning efficiently and maintain performance, especially at high sparsity ratios. To this end, we introduce an efficient structured pruning framework named CFSP, which leverages both Coarse (interblock) and Fine-grained (intrablock) activation information as an importance criterion to guide pruning. The pruning is highly efficient, as it only requires one forward pass to compute feature activations. Specifically, we first allocate the sparsity budget across blocks based on their importance and then retain important weights within each block. In addition, we introduce a recovery fine-tuning strategy that adaptively allocates training overhead based on coarse-grained importance to further improve performance. Experimental results demonstrate that CFSP outperforms existing methods on diverse models across various sparsity budgets. Our code will be available at https://github.com/wyxscir/CFSP.
comment: Proc. The 31st International Conference on Computational Linguistics (COLING2025)
♻ ☆ Automating Governing Knowledge Commons and Contextual Integrity (GKC-CI) Privacy Policy Annotations with Large Language Models
Identifying contextual integrity (CI) and governing knowledge commons (GKC) parameters in privacy policy texts can facilitate normative privacy analysis. However, GKC-CI annotation has heretofore required manual or crowdsourced effort. This paper demonstrates that high-accuracy GKC-CI parameter annotation of privacy policies can be performed automatically using large language models. We fine-tune 50 open-source and proprietary models on 21,588 ground truth GKC-CI annotations from 16 privacy policies. Our best performing model has an accuracy of 90.65%, which is comparable to the accuracy of experts on the same task. We apply our best performing model to 456 privacy policies from a variety of online services, demonstrating the effectiveness of scaling GKC-CI annotation for privacy policy exploration and analysis. We publicly release our model training code, training and testing data, an annotation visualizer, and all annotated policies for future GKC-CI research.
comment: 29 pages, 18 figures, 11 tables; camera-ready version
♻ ☆ Will sentiment analysis need subculture? A new data augmentation approach
Nowadays, the omnipresence of the Internet has fostered a subculture that congregates around the contemporary milieu. The subculture artfully articulates the intricacies of human feelings by ardently pursuing the allure of novelty, a fact that cannot be disregarded in the sentiment analysis. This paper aims to enrich data through the lens of subculture, to address the insufficient training data faced by sentiment analysis. To this end, a new approach of subculture-based data augmentation (SCDA) is proposed, which engenders enhanced texts for each training text by leveraging the creation of specific subcultural expression generators. The extensive experiments attest to the effectiveness and potential of SCDA. The results also shed light on the phenomenon that disparate subcultural expressions elicit varying degrees of sentiment stimulation. Moreover, an intriguing conjecture arises, suggesting the linear reversibility of certain subcultural expressions.
comment: JASIST
♻ ☆ BnSentMix: A Diverse Bengali-English Code-Mixed Dataset for Sentiment Analysis COLING 2025
The widespread availability of code-mixed data can provide valuable insights into low-resource languages like Bengali, which have limited datasets. Sentiment analysis has been a fundamental text classification task across several languages for code-mixed data. However, there has yet to be a large-scale and diverse sentiment analysis dataset on code-mixed Bengali. We address this limitation by introducing BnSentMix, a sentiment analysis dataset on code-mixed Bengali consisting of 20,000 samples with 4 sentiment labels from Facebook, YouTube, and e-commerce sites. We ensure diversity in data sources to replicate realistic code-mixed scenarios. Additionally, we propose 14 baseline methods including novel transformer encoders further pre-trained on code-mixed Bengali-English, achieving an overall accuracy of 69.8% and an F1 score of 69.1% on sentiment classification tasks. Detailed analyses reveal variations in performance across different sentiment labels and text types, highlighting areas for future improvement.
comment: LoResLM at COLING 2025
♻ ☆ Interpretable Company Similarity with Sparse Autoencoders
Determining company similarity is a vital task in finance, underpinning hedging, risk management, portfolio diversification, and more. Practitioners often rely on sector and industry classifications to gauge similarity, such as SIC-codes and GICS-codes - the former being used by the U.S. Securities and Exchange Commission (SEC), and the latter widely used by the investment community. Since these classifications can lack granularity and often need to be updated, using clusters of embeddings of company descriptions has been proposed as a potential alternative, but the lack of interpretability in token embeddings poses a significant barrier to adoption in high-stakes contexts. Sparse Autoencoders (SAEs) have shown promise in enhancing the interpretability of Large Language Models (LLMs) by decomposing LLM activations into interpretable features. We apply SAEs to company descriptions, obtaining meaningful clusters of equities in the process. We benchmark SAE features against SIC-codes, Major Group codes, and Embeddings. Our results demonstrate that SAE features not only replicate but often surpass sector classifications and embeddings in capturing fundamental company characteristics. This is evidenced by their superior performance in correlating monthly returns - a proxy for similarity - and generating higher Sharpe ratio co-integration strategies, which underscores deeper fundamental similarities among companies.
♻ ☆ Merino: Entropy-driven Design for Generative Language Models on IoT Devices AAAI 2025
Generative Large Language Models (LLMs) stand as a revolutionary advancement in the modern era of artificial intelligence (AI). However, scaling down LLMs for resource-constrained hardware, such as Internet-of-Things (IoT) devices requires non-trivial efforts and domain knowledge. In this paper, we propose a novel information-entropy framework for designing mobile-friendly generative language models. The whole design procedure involves solving a mathematical programming (MP) problem, which can be done on the CPU within minutes, making it nearly zero-cost. We evaluate our designed models, termed MeRino, across fourteen NLP downstream tasks, showing their competitive performance against the state-of-the-art autoregressive transformer models under the mobile setting. Notably, MeRino achieves similar or better performance on both language modeling and zero-shot learning tasks, compared to the 350M parameter OPT while being 4.9x faster on NVIDIA Jetson Nano with 5.5x reduction in model size.
comment: AAAI 2025
♻ ☆ LLaSA: A Multimodal LLM for Human Activity Analysis Through Wearable and Smartphone Sensors
Integrating inertial measurement units (IMUs) with large language models (LLMs) expands the potential of multimodal AI, enabling more nuanced human activity analysis. In this paper, we introduce LLaSA (Large Language and Sensor Assistant), a multimodal large language model built on LIMU-BERT and Llama, designed to interpret and answer queries related to human activities and motion analysis, leveraging sensor data and contextual reasoning. To develop LLaSA, we introduce two key datasets: SensorCaps, a comprehensive collection of 35,960 IMU-derived narratives with handcrafted features, and OpenSQA, an instruction-following dataset containing 179,727 question-answer pairs aware of the sensor and human activity context. These datasets provide diverse and rich inputs to train LLaSA for complex sensor-based queries. To optimize LLaSA's performance, we apply a unique hyperparameter tuning method, which significantly enhances its effectiveness in contextual question-answering tasks. Extensive evaluations, including a human-led assessment of the question-answering, demonstrate that LLaSA achieves superior data interpretation and context-aware responses compared to GPT-3.5-Turbo and Vicuna-1.5-13b-16K. These contributions advance the frontier of sensor-aware LLMs and create new opportunities for impactful multimodal research in healthcare, sports science, and human-computer interactions. Our code repository and datasets can be found at https://github.com/BASHLab/LLaSA.
comment: Under review at SenSys 2025
♻ ☆ Exploring Knowledge Tracing in Tutor-Student Dialogues using LLMs
Recent advances in large language models (LLMs) have led to the development of artificial intelligence (AI)-powered tutoring chatbots, showing promise in providing broad access to high-quality personalized education. Existing works have studied how to make LLMs follow tutoring principles, but have not studied broader uses of LLMs for supporting tutoring. Up until now, tracing student knowledge and analyzing misconceptions has been difficult and time-consuming to implement for open-ended dialogue tutoring. In this work, we investigate whether LLMs can be supportive of this task: we first use LLM prompting methods to identify the knowledge components/skills involved in each dialogue turn, i.e., a tutor utterance posing a task or a student utterance that responds to it. We also evaluate whether the student responds correctly to the tutor and verify the LLM's accuracy using human expert annotations. We then apply a range of knowledge tracing (KT) methods on the resulting labeled data to track student knowledge levels over an entire dialogue. We conduct experiments on two tutoring dialogue datasets, and show that a novel yet simple LLM-based method, LLMKT, significantly outperforms existing KT methods in predicting student response correctness in dialogues. We perform extensive qualitative analyses to highlight the challenges in dialogueKT and outline multiple avenues for future work.
comment: Published in LAK 2025: The 15th International Learning Analytics and Knowledge Conference
♻ ☆ Summon a Demon and Bind it: A Grounded Theory of LLM Red Teaming
Engaging in the deliberate generation of abnormal outputs from Large Language Models (LLMs) by attacking them is a novel human activity. This paper presents a thorough exposition of how and why people perform such attacks, defining LLM red-teaming based on extensive and diverse evidence. Using a formal qualitative methodology, we interviewed dozens of practitioners from a broad range of backgrounds, all contributors to this novel work of attempting to cause LLMs to fail. We focused on the research questions of defining LLM red teaming, uncovering the motivations and goals for performing the activity, and characterizing the strategies people use when attacking LLMs. Based on the data, LLM red teaming is defined as a limit-seeking, non-malicious, manual activity, which depends highly on a team-effort and an alchemist mindset. It is highly intrinsically motivated by curiosity, fun, and to some degrees by concerns for various harms of deploying LLMs. We identify a taxonomy of 12 strategies and 35 different techniques of attacking LLMs. These findings are presented as a comprehensive grounded theory of how and why people attack large language models: LLM red teaming.
♻ ☆ BiMediX: Bilingual Medical Mixture of Experts LLM EMNLP 2024
In this paper, we introduce BiMediX, the first bilingual medical mixture of experts LLM designed for seamless interaction in both English and Arabic. Our model facilitates a wide range of medical interactions in English and Arabic, including multi-turn chats to inquire about additional details such as patient symptoms and medical history, multiple-choice question answering, and open-ended question answering. We propose a semi-automated English-to-Arabic translation pipeline with human refinement to ensure high-quality translations. We also introduce a comprehensive evaluation benchmark for Arabic medical LLMs. Furthermore, we introduce BiMed1.3M, an extensive Arabic-English bilingual instruction set covering 1.3 Million diverse medical interactions, resulting in over 632 million healthcare specialized tokens for instruction tuning. Our BiMed1.3M dataset includes 250k synthesized multi-turn doctor-patient chats and maintains a 1:2 Arabic-to-English ratio. Our model outperforms state-of-the-art Med42 and Meditron by average absolute gains of 2.5% and 4.1%, respectively, computed across multiple medical evaluation benchmarks in English, while operating at 8-times faster inference. Moreover, our BiMediX outperforms the generic Arabic-English bilingual LLM, Jais-30B, by average absolute gains of 10% on our Arabic medical benchmark and 15% on bilingual evaluations across multiple datasets. Our project page with source code and trained model is available at https://github.com/mbzuai-oryx/BiMediX .
comment: Accepted to EMNLP 2024 (Findings)
♻ ☆ WSI-LLaVA: A Multimodal Large Language Model for Whole Slide Image
Recent advancements in computational pathology have produced patch-level Multi-modal Large Language Models (MLLMs), but these models are limited by their inability to analyze whole slide images (WSIs) comprehensively and their tendency to bypass crucial morphological features that pathologists rely on for diagnosis. To address these challenges, we first introduce WSI-Bench, a large-scale morphology-aware benchmark containing 180k VQA pairs from 9,850 WSIs across 30 cancer types, designed to evaluate MLLMs' understanding of morphological characteristics crucial for accurate diagnosis. Building upon this benchmark, we present WSI-LLaVA, a novel framework for gigapixel WSI understanding that employs a three-stage training approach: WSI-text alignment, feature space alignment, and task-specific instruction tuning. To better assess model performance in pathological contexts, we develop two specialized WSI metrics: WSI-Precision and WSI-Relevance. Experimental results demonstrate that WSI-LLaVA outperforms existing models across all capability dimensions, with a significant improvement in morphological analysis, establishing a clear correlation between morphological understanding and diagnostic accuracy.
comment: 38 pages, 22 figures, 35 tables
♻ ☆ MATATA: A weakly-supervised MAthematical Tool-Assisted reasoning for Tabular Applications
Mathematical reasoning capabilities are increasing with tool-augmented language agents, but methods often rely either on closed-source or large models, external data, or extensive prompt engineering. This work introduces MATATA, a novel cost-effective method to train LLM agents for tabular data problems through reasoning, planning, and tool use. With a progressive self-improvement paradigm and an iterative weak supervision, it empowers 3.8B/8B Small Language Models (SLMs), particularly suited for local hosting and sensitive business contexts where data privacy is crucial. By employing a flexible and reusable tools across different datasets, it achieves robust performance with effective scalability across shared tasks. Experiments show that MATATA reaches state-of-the-art performances on FinQA and TAT-QA among reasoning frameworks based on open-source models. Moreover, MATATA models compete with GPT-4 based frameworks on TabMWP, while being SLMs.
Information Retrieval 21
☆ Benchmark for Evaluation and Analysis of Citation Recommendation Models
Citation recommendation systems have attracted much academic interest, resulting in many studies and implementations. These systems help authors automatically generate proper citations by suggesting relevant references based on the text they have written. However, the methods used in citation recommendation differ across various studies and implementations. Some approaches focus on the overall content of papers, while others consider the context of the citation text. Additionally, the datasets used in these studies include different aspects of papers, such as metadata, citation context, or even the full text of the paper in various formats and structures. The diversity in models, datasets, and evaluation metrics makes it challenging to assess and compare citation recommendation methods effectively. To address this issue, a standardized dataset and evaluation metrics are needed to evaluate these models consistently. Therefore, we propose developing a benchmark specifically designed to analyze and compare citation recommendation models. This benchmark will evaluate the performance of models on different features of the citation context and provide a comprehensive evaluation of the models across all these tasks, presenting the results in a standardized way. By creating a benchmark with standardized evaluation metrics, researchers and practitioners in the field of citation recommendation will have a common platform to assess and compare different models. This will enable meaningful comparisons and help identify promising approaches for further research and development in the field.
comment: 10 pages
☆ OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations
Document content extraction is crucial in computer vision, especially for meeting the high-quality data needs of large language models (LLMs) and retrieval-augmented generation (RAG) technologies. However, current document parsing methods suffer from significant limitations in terms of diversity and comprehensive evaluation. To address these challenges, we introduce OmniDocBench, a novel multi-source benchmark designed to advance automated document content extraction. OmniDocBench includes a meticulously curated and annotated high-quality evaluation dataset comprising nine diverse document types, such as academic papers, textbooks, slides, among others. Our benchmark provides a flexible and comprehensive evaluation framework with 19 layout category labels and 14 attribute labels, enabling multi-level assessments across entire datasets, individual modules, or specific data types. Using OmniDocBench, we perform an exhaustive comparative analysis of existing modular pipelines and multimodal end-to-end methods, highlighting their limitations in handling document diversity and ensuring fair evaluation. OmniDocBench establishes a robust, diverse, and fair evaluation standard for the document content extraction field, offering crucial insights for future advancements and fostering the development of document parsing technologies. The codes and dataset is available in https://github.com/opendatalab/OmniDocBench.
☆ SST framework for Document Matching
Long-form document matching aims to judge the relevance between two documents and has been applied to various scenarios. Most existing works utilize hierarchical or long context models to process documents, which achieve coarse understanding but may ignore details. Some researchers construct a document view with similar sentences about aligned document subtopics to focus on detailed matching signals. However, a long document generally contains multiple subtopics. The matching signals are heterogeneous from multiple topics. Considering only the homologous aligned subtopics may not be representative enough and may cause biased modeling. In this paper, we introduce a new framework to model representative matching signals. First, we propose to capture various matching signals through subtopics of document pairs. Next, We construct multiple document views based on subtopics to cover heterogeneous and valuable details. However, existing spatial aggregation methods like attention, which integrate all these views simultaneously, are hard to integrate heterogeneous information. Instead, we propose temporal aggregation, which effectively integrates different views gradually as the training progresses. Experimental results show that our learning framework is effective on several document-matching tasks, including news duplication and legal case retrieval.
☆ Bilingual BSARD: Extending Statutory Article Retrieval to Dutch COLING
Statutory article retrieval plays a crucial role in making legal information more accessible to both laypeople and legal professionals. Multilingual countries like Belgium present unique challenges for retrieval models due to the need for handling legal issues in multiple languages. Building on the Belgian Statutory Article Retrieval Dataset (BSARD) in French, we introduce the bilingual version of this dataset, bBSARD. The dataset contains parallel Belgian statutory articles in both French and Dutch, along with legal questions from BSARD and their Dutch translation. Using bBSARD, we conduct extensive benchmarking of retrieval models available for Dutch and French. Our benchmarking setup includes lexical models, zero-shot dense models, and fine-tuned small foundation models. Our experiments show that BM25 remains a competitive baseline compared to many zero-shot dense models in both languages. We also observe that while proprietary models outperform open alternatives in the zero-shot setting, they can be matched or surpassed by fine-tuning small language-specific models. Our dataset and evaluation code are publicly available.
comment: To be presented at RegNLP-2025 (COLING)
☆ RAG-based Question Answering over Heterogeneous Data and Text
This article presents the QUASAR system for question answering over unstructured text, structured tables, and knowledge graphs, with unified treatment of all sources. The system adopts a RAG-based architecture, with a pipeline of evidence retrieval followed by answer generation, with the latter powered by a moderate-sized language model. Additionally and uniquely, QUASAR has components for question understanding, to derive crisper input for evidence retrieval, and for re-ranking and filtering the retrieved evidence before feeding the most informative pieces into the answer generation. Experiments with three different benchmarks demonstrate the high answering quality of our approach, being on par with or better than large GPT models, while keeping the computational cost and energy consumption orders of magnitude lower.
comment: IEEE Data Engineering Bulletin -- December 2024 Edition on RAG
☆ RLT4Rec: Reinforcement Learning Transformer for User Cold Start and Item Recommendation
We introduce a new sequential transformer reinforcement learning architecture RLT4Rec and demonstrate that it achieves excellent performance in a range of item recommendation tasks. RLT4Rec uses a relatively simple transformer architecture that takes as input the user's (item,rating) history and outputs the next item to present to the user. Unlike existing RL approaches, there is no need to input a state observation or estimate. RLT4Rec handles new users and established users within the same consistent framework and automatically balances the "exploration" needed to discover the preferences of a new user with the "exploitation" that is more appropriate for established users. Training of RLT4Rec is robust and fast and is insensitive to the choice of training data, learning to generate "good" personalised sequences that the user tends to rate highly even when trained on "bad" data.
☆ Temporal Linear Item-Item Model for Sequential Recommendation WSDM 2025
In sequential recommendation (SR), neural models have been actively explored due to their remarkable performance, but they suffer from inefficiency inherent to their complexity. On the other hand, linear SR models exhibit high efficiency and achieve competitive or superior accuracy compared to neural models. However, they solely deal with the sequential order of items (i.e., sequential information) and overlook the actual timestamp (i.e., temporal information). It is limited to effectively capturing various user preference drifts over time. To address this issue, we propose a novel linear SR model, named TemporAl LinEar item-item model (TALE), incorporating temporal information while preserving training/inference efficiency, with three key components. (i) Single-target augmentation concentrates on a single target item, enabling us to learn the temporal correlation for the target item. (ii) Time interval-aware weighting utilizes the actual timestamp to discern the item correlation depending on time intervals. (iii) Trend-aware normalization reflects the dynamic shift of item popularity over time. Our empirical studies show that TALE outperforms ten competing SR models by up to 18.71% gains on five benchmark datasets. It also exhibits remarkable effectiveness in evaluating long-tail items by up to 30.45% gains. The source code is available at https://github.com/psm1206/TALE.
comment: Accepted by WSDM 2025
☆ IntellectSeeker: A Personalized Literature Management System with the Probabilistic Model and Large Language Model
Faced with the burgeoning volume of academic literature, researchers often need help with uncertain article quality and mismatches in term searches using traditional academic engines. We introduce IntellectSeeker, an innovative and personalized intelligent academic literature management platform to address these challenges. This platform integrates a Large Language Model (LLM)--based semantic enhancement bot with a sophisticated probability model to personalize and streamline literature searches. We adopted the GPT-3.5-turbo model to transform everyday language into professional academic terms across various scenarios using multiple rounds of few-shot learning. This adaptation mainly benefits academic newcomers, effectively bridging the gap between general inquiries and academic terminology. The probabilistic model intelligently filters academic articles to align closely with the specific interests of users, which are derived from explicit needs and behavioral patterns. Moreover, IntellectSeeker incorporates an advanced recommendation system and text compression tools. These features enable intelligent article recommendations based on user interactions and present search results through concise one-line summaries and innovative word cloud visualizations, significantly enhancing research efficiency and user experience. IntellectSeeker offers academic researchers a highly customizable literature management solution with exceptional search precision and matching capabilities. The code can be found here: https://github.com/LuckyBian/ISY5001
☆ Combining knowledge graphs and LLMs for hazardous chemical information management and reuse
Human health is increasingly threatened by exposure to hazardous substances, particularly persistent and toxic chemicals. The link between these substances, often encountered in complex mixtures, and various diseases are demonstrated in scientific studies. However, this information is scattered across several sources and hardly accessible by humans and machines. This paper evaluates current practices for publishing/accessing information on hazardous chemicals and proposes a novel platform designed to facilitate retrieval of critical chemical data in urgent situations. The platform aggregates information from multiple sources and organizes it into a structured knowledge graph. Users can access this information through a visual interface such as Neo4J Bloom and dashboards, or via natural language queries using a Chatbot. Our findings demonstrate a significant reduction in the time and effort required to access vital chemical information when datasets follow FAIR principles. Furthermore, we discuss the lessons learned from the development and implementation of this platform and provide recommendations for data owners and publishers to enhance data reuse and interoperability. This work aims to improve the accessibility and usability of chemical information by healthcare professionals, thereby supporting better health outcomes and informed decision-making in the face of patients exposed to chemical intoxication risks.
comment: Submitted to IEEE BIBM24
♻ ☆ Beyond Retrieval: Generating Narratives in Conversational Recommender Systems
The recent advances in Large Language Model's generation and reasoning capabilities present an opportunity to develop truly conversational recommendation systems. However, effectively integrating recommender system knowledge into LLMs for natural language generation which is tailored towards recommendation tasks remains a challenge. This paper addresses this challenge by making two key contributions. First, we introduce a new dataset (REGEN) for natural language generation tasks in conversational recommendations. REGEN (Reviews Enhanced with GEnerative Narratives) extends the Amazon Product Reviews dataset with rich user narratives, including personalized explanations of product preferences, product endorsements for recommended items, and summaries of user purchase history. REGEN is made publicly available to facilitate further research. Furthermore, we establish benchmarks using well-known generative metrics, and perform an automated evaluation of the new dataset using a rater LLM. Second, the paper introduces a fusion architecture (CF model with an LLM) which serves as a baseline for REGEN. And to the best of our knowledge, represents the first attempt to analyze the capabilities of LLMs in understanding recommender signals and generating rich narratives. We demonstrate that LLMs can effectively learn from simple fusion architectures utilizing interaction-based CF embeddings, and this can be further enhanced using the metadata and personalization data associated with items. Our experiments show that combining CF and content embeddings leads to improvements of 4-12% in key language metrics compared to using either type of embedding individually. We also provide an analysis to interpret how CF and content embeddings contribute to this new generative task.
♻ ☆ Optimistic Query Routing in Clustering-based Approximate Maximum Inner Product Search
Clustering-based nearest neighbor search is an effective method in which points are partitioned into geometric shards to form an index, with only a few shards searched during query processing to find a set of top-$k$ vectors. Even though the search efficacy is heavily influenced by the algorithm that identifies the shards to probe, it has received little attention in the literature. This work bridges that gap by studying routing in clustering-based maximum inner product search. We unpack existing routers and notice the surprising contribution of optimism. We then take a page from the sequential decision making literature and formalize that insight following the principle of ``optimism in the face of uncertainty.'' In particular, we present a framework that incorporates the moments of the distribution of inner products within each shard to estimate the maximum inner product. We then present an instance of our algorithm that uses only the first two moments to reach the same accuracy as state-of-the-art routers such as ScaNN by probing up to $50\%$ fewer points on benchmark datasets. Our algorithm is also space-efficient: we design a sketch of the second moment whose size is independent of the number of points and requires $\mathcal{O}(1)$ vectors per shard.
♻ ☆ CFaiRLLM: Consumer Fairness Evaluation in Large-Language Model Recommender System
This work takes a critical stance on previous studies concerning fairness evaluation in Large Language Model (LLM)-based recommender systems, which have primarily assessed consumer fairness by comparing recommendation lists generated with and without sensitive user attributes. Such approaches implicitly treat discrepancies in recommended items as biases, overlooking whether these changes might stem from genuine personalization aligned with true preferences of users. Moreover, these earlier studies typically address single sensitive attributes in isolation, neglecting the complex interplay of intersectional identities. In response to these shortcomings, we introduce CFaiRLLM, an enhanced evaluation framework that not only incorporates true preference alignment but also rigorously examines intersectional fairness by considering overlapping sensitive attributes. Additionally, CFaiRLLM introduces diverse user profile sampling strategies-random, top-rated, and recency-focused-to better understand the impact of profile generation fed to LLMs in light of inherent token limitations in these systems. Given that fairness depends on accurately understanding users' tastes and preferences,, these strategies provide a more realistic assessment of fairness within RecLLMs. The results demonstrated that true preference alignment offers a more personalized and fair assessment compared to similarity-based measures, revealing significant disparities when sensitive and intersectional attributes are incorporated. Notably, our study finds that intersectional attributes amplify fairness gaps more prominently, especially in less structured domains such as music recommendations in LastFM.
♻ ☆ S+t-SNE -- Bringing Dimensionality Reduction to Data Streams
We present S+t-SNE, an adaptation of the t-SNE algorithm designed to handle infinite data streams. The core idea behind S+t-SNE is to update the t-SNE embedding incrementally as new data arrives, ensuring scalability and adaptability to handle streaming scenarios. By selecting the most important points at each step, the algorithm ensures scalability while keeping informative visualisations. By employing a blind method for drift management, the algorithm adjusts the embedding space, which facilitates the visualisation of evolving data dynamics. Our experimental evaluations demonstrate the effectiveness and efficiency of S+t-SNE, whilst highlighting its ability to capture patterns in a streaming scenario. We hope our approach offers researchers and practitioners a real-time tool for understanding and interpreting high-dimensional data.
comment: This preprint has undergone peer review but does not have any post-submission improvements or corrections. Full version after peer-review and post-acceptance improvements was presented at IDA2024 (https://ida2024.org/)
♻ ☆ Hypergrah-Enhanced Dual Convolutional Network for Bundle Recommendation
Bundle recommendations strive to offer users a set of items as a package named bundle, enhancing convenience and contributing to the seller's revenue. While previous approaches have demonstrated notable performance, we argue that they may compromise the ternary relationship among users, items, and bundles. This compromise can result in information loss, ultimately impacting the overall model performance. To address this gap, we develop a unified model for bundle recommendation, termed hypergraph-enhanced dual convolutional neural network (HED). Our approach is characterized by two key aspects. Firstly, we construct a complete hypergraph to capture interaction dynamics among users, items, and bundles. Secondly, we incorporate U-B interaction information to enhance the information representation derived from users and bundle embedding vectors. Extensive experimental results on the Youshu and Netease datasets have demonstrated that HED surpasses state-of-the-art baselines, proving its effectiveness. In addition, various ablation studies and sensitivity analyses revealed the working mechanism and proved our effectiveness. Codes and datasets are available at https://github.com/AAI-Lab/HED
♻ ☆ Causal Deconfounding via Confounder Disentanglement for Dual-Target Cross-Domain Recommendation
In recent years, dual-target Cross-Domain Recommendation (CDR) has been proposed to capture comprehensive user preferences in order to ultimately enhance the recommendation accuracy in both data-richer and data-sparser domains simultaneously. However, in addition to users' true preferences, the user-item interactions might also be affected by confounders (e.g., free shipping, sales promotion). As a result, dual-target CDR has to meet two challenges: (1) how to effectively decouple observed confounders, including single-domain confounders and cross-domain confounders, and (2) how to preserve the positive effects of observed confounders on predicted interactions, while eliminating their negative effects on capturing comprehensive user preferences. To address the above two challenges, we propose a Causal Deconfounding framework via Confounder Disentanglement for dual-target Cross-Domain Recommendation, called CD2CDR. In CD2CDR, we first propose a confounder disentanglement module to effectively decouple observed single-domain and cross-domain confounders. We then propose a causal deconfounding module to preserve the positive effects of such observed confounders and eliminate their negative effects via backdoor adjustment, thereby enhancing the recommendation accuracy in each domain. Extensive experiments conducted on five real-world datasets demonstrate that CD2CDR significantly outperforms the state-of-the-art methods.
♻ ☆ Argumentative Experience: Reducing Confirmation Bias on Controversial Issues through LLM-Generated Multi-Persona Debates
Large language models (LLMs) are enabling designers to give life to exciting new user experiences for information access. In this work, we present a system that generates LLM personas to debate a topic of interest from different perspectives. How might information seekers use and benefit from such a system? Can centering information access around diverse viewpoints help to mitigate thorny challenges like confirmation bias in which information seekers over-trust search results matching existing beliefs? How do potential biases and hallucinations in LLMs play out alongside human users who are also fallible and possibly biased? Our study exposes participants to multiple viewpoints on controversial issues via a mixed-methods, within-subjects study. We use eye-tracking metrics to quantitatively assess cognitive engagement alongside qualitative feedback. Compared to a baseline search system, we see more creative interactions and diverse information-seeking with our multi-persona debate system, which more effectively reduces user confirmation bias and conviction toward their initial beliefs. Overall, our study contributes to the emerging design space of LLM-based information access systems, specifically investigating the potential of simulated personas to promote greater exposure to information diversity, emulate collective intelligence, and mitigate bias in information seeking.
♻ ☆ LEARN: Knowledge Adaptation from Large Language Model to Recommendation for Practical Industrial Application AAAI 2025
Contemporary recommendation systems predominantly rely on ID embedding to capture latent associations among users and items. However, this approach overlooks the wealth of semantic information embedded within textual descriptions of items, leading to suboptimal performance and poor generalizations. Leveraging the capability of large language models to comprehend and reason about textual content presents a promising avenue for advancing recommendation systems. To achieve this, we propose an Llm-driven knowlEdge Adaptive RecommeNdation (LEARN) framework that synergizes open-world knowledge with collaborative knowledge. We address computational complexity concerns by utilizing pretrained LLMs as item encoders and freezing LLM parameters to avoid catastrophic forgetting and preserve open-world knowledge. To bridge the gap between the open-world and collaborative domains, we design a twin-tower structure supervised by the recommendation task and tailored for practical industrial application. Through experiments on the real large-scale industrial dataset and online A/B tests, we demonstrate the efficacy of our approach in industry application. We also achieve state-of-the-art performance on six Amazon Review datasets to verify the superiority of our method.
comment: Accepted by AAAI 2025
♻ ☆ LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods
The rapid advancement of Large Language Models (LLMs) has driven their expanding application across various fields. One of the most promising applications is their role as evaluators based on natural language responses, referred to as ''LLMs-as-judges''. This framework has attracted growing attention from both academia and industry due to their excellent effectiveness, ability to generalize across tasks, and interpretability in the form of natural language. This paper presents a comprehensive survey of the LLMs-as-judges paradigm from five key perspectives: Functionality, Methodology, Applications, Meta-evaluation, and Limitations. We begin by providing a systematic definition of LLMs-as-Judges and introduce their functionality (Why use LLM judges?). Then we address methodology to construct an evaluation system with LLMs (How to use LLM judges?). Additionally, we investigate the potential domains for their application (Where to use LLM judges?) and discuss methods for evaluating them in various contexts (How to evaluate LLM judges?). Finally, we provide a detailed analysis of the limitations of LLM judges and discuss potential future directions. Through a structured and comprehensive analysis, we aim aims to provide insights on the development and application of LLMs-as-judges in both research and practice. We will continue to maintain the relevant resource list at https://github.com/CSHaitao/Awesome-LLMs-as-Judges.
comment: 60 pages, comprehensive and continuously updated
♻ ☆ Personalized Federated Collaborative Filtering: A Variational AutoEncoder Approach
Federated Collaborative Filtering (FedCF) is an emerging field focused on developing a new recommendation framework with preserving privacy in a federated setting. Existing FedCF methods typically combine distributed Collaborative Filtering (CF) algorithms with privacy-preserving mechanisms, and then preserve personalized information into a user embedding vector. However, the user embedding is usually insufficient to preserve the rich information of the fine-grained personalization across heterogeneous clients. This paper proposes a novel personalized FedCF method by preserving users' personalized information into a latent variable and a neural model simultaneously. Specifically, we decompose the modeling of user knowledge into two encoders, each designed to capture shared knowledge and personalized knowledge separately. A personalized gating network is then applied to balance personalization and generalization between the global and local encoders. Moreover, to effectively train the proposed framework, we model the CF problem as a specialized Variational AutoEncoder (VAE) task by integrating user interaction vector reconstruction with missing value prediction. The decoder is trained to reconstruct the implicit feedback from items the user has interacted with, while also predicting items the user might be interested in but has not yet interacted with. Experimental results on benchmark datasets demonstrate that the proposed method outperforms other baseline methods, showcasing superior performance. Our code is available at https://github.com/mtics/FedDAE.
comment: 10 pages, 3 figures, 4 tables, conference
♻ ☆ BayesCNS: A Unified Bayesian Approach to Address Cold Start and Non-Stationarity in Search Systems at Scale
Information Retrieval (IR) systems used in search and recommendation platforms frequently employ Learning-to-Rank (LTR) models to rank items in response to user queries. These models heavily rely on features derived from user interactions, such as clicks and engagement data. This dependence introduces cold start issues for items lacking user engagement and poses challenges in adapting to non-stationary shifts in user behavior over time. We address both challenges holistically as an online learning problem and propose BayesCNS, a Bayesian approach designed to handle cold start and non-stationary distribution shifts in search systems at scale. BayesCNS achieves this by estimating prior distributions for user-item interactions, which are continuously updated with new user interactions gathered online. This online learning procedure is guided by a ranker model, enabling efficient exploration of relevant items using contextual information provided by the ranker. We successfully deployed BayesCNS in a large-scale search system and demonstrated its efficacy through comprehensive offline and online experiments. Notably, an online A/B experiment showed a 10.60% increase in new item interactions and a 1.05% improvement in overall success metrics over the existing production baseline.
♻ ☆ MARec: Metadata Alignment for cold-start Recommendation
For many recommender systems, the primary data source is a historical record of user clicks. The associated click matrix is often very sparse, as the number of users x products can be far larger than the number of clicks. Such sparsity is accentuated in cold-start settings, which makes the efficient use of metadata information of paramount importance. In this work, we propose a simple approach to address cold-start recommendations by leveraging content metadata, Metadata Alignment for cold-start Recommendation. We show that this approach can readily augment existing matrix factorization and autoencoder approaches, enabling a smooth transition to top performing algorithms in warmer set-ups. Our experimental results indicate three separate contributions: first, we show that our proposed framework largely beats SOTA results on 4 cold-start datasets with different sparsity and scale characteristics, with gains ranging from +8.4% to +53.8% on reported ranking metrics; second, we provide an ablation study on the utility of semantic features, and proves the additional gain obtained by leveraging such features ranges between +46.8% and +105.5%; and third, our approach is by construction highly competitive in warm set-ups, and we propose a closed-form solution outperformed by SOTA results by only 0.8% on average.
Machine Learning 150
Efficient Diversity-Preserving Diffusion Alignment via Gradient-Informed GFlowNets
While one commonly trains large diffusion models by collecting datasets on target downstream tasks, it is often desired to align and finetune pretrained diffusion models on some reward functions that are either designed by experts or learned from small-scale datasets. Existing methods for finetuning diffusion models typically suffer from lack of diversity in generated samples, lack of prior preservation, and/or slow convergence in finetuning. Inspired by recent successes in generative flow networks (GFlowNets), a class of probabilistic models that sample with the unnormalized density of a reward function, we propose a novel GFlowNet method dubbed Nabla-GFlowNet (abbreviated as $\nabla$-GFlowNet), the first GFlowNet method that leverages the rich signal in reward gradients, together with an objective called $\nabla$-DB plus its variant residual $\nabla$-DB designed for prior-preserving diffusion alignment. We show that our proposed method achieves fast yet diversity- and prior-preserving alignment of Stable Diffusion, a large-scale text-conditioned image diffusion model, on different realistic reward functions.
comment: Technical Report (35 pages, 31 figures)
☆ Video Motion Transfer with Diffusion Transformers
We propose DiTFlow, a method for transferring the motion of a reference video to a newly synthesized one, designed specifically for Diffusion Transformers (DiT). We first process the reference video with a pre-trained DiT to analyze cross-frame attention maps and extract a patch-wise motion signal called the Attention Motion Flow (AMF). We guide the latent denoising process in an optimization-based, training-free, manner by optimizing latents with our AMF loss to generate videos reproducing the motion of the reference one. We also apply our optimization strategy to transformer positional embeddings, granting us a boost in zero-shot motion transfer capabilities. We evaluate DiTFlow against recently published methods, outperforming all across multiple metrics and human evaluation.
comment: Project page: https://ditflow.github.io/
☆ Mobile-TeleVision: Predictive Motion Priors for Humanoid Whole-Body Control
Humanoid robots require both robust lower-body locomotion and precise upper-body manipulation. While recent Reinforcement Learning (RL) approaches provide whole-body loco-manipulation policies, they lack precise manipulation with high DoF arms. In this paper, we propose decoupling upper-body control from locomotion, using inverse kinematics (IK) and motion retargeting for precise manipulation, while RL focuses on robust lower-body locomotion. We introduce PMP (Predictive Motion Priors), trained with Conditional Variational Autoencoder (CVAE) to effectively represent upper-body motions. The locomotion policy is trained conditioned on this upper-body motion representation, ensuring that the system remains robust with both manipulation and locomotion. We show that CVAE features are crucial for stability and robustness, and significantly outperforms RL-based whole-body control in precise manipulation. With precise upper-body motion and robust lower-body locomotion control, operators can remotely control the humanoid to walk around and explore different environments, while performing diverse manipulation tasks.
☆ From an Image to a Scene: Learning to Imagine the World from a Million 360 Videos NeurIPS 2024
Three-dimensional (3D) understanding of objects and scenes play a key role in humans' ability to interact with the world and has been an active area of research in computer vision, graphics, and robotics. Large scale synthetic and object-centric 3D datasets have shown to be effective in training models that have 3D understanding of objects. However, applying a similar approach to real-world objects and scenes is difficult due to a lack of large-scale data. Videos are a potential source for real-world 3D data, but finding diverse yet corresponding views of the same content has shown to be difficult at scale. Furthermore, standard videos come with fixed viewpoints, determined at the time of capture. This restricts the ability to access scenes from a variety of more diverse and potentially useful perspectives. We argue that large scale 360 videos can address these limitations to provide: scalable corresponding frames from diverse views. In this paper, we introduce 360-1M, a 360 video dataset, and a process for efficiently finding corresponding frames from diverse viewpoints at scale. We train our diffusion-based model, Odin, on 360-1M. Empowered by the largest real-world, multi-view dataset to date, Odin is able to freely generate novel views of real-world scenes. Unlike previous methods, Odin can move the camera through the environment, enabling the model to infer the geometry and layout of the scene. Additionally, we show improved performance on standard novel view synthesis and 3D reconstruction benchmarks.
comment: NeurIPS 2024. For project page, see https://mattwallingford.github.io/ODIN
☆ Bayesian Optimization of Antibodies Informed by a Generative Model of Evolving Sequences
To build effective therapeutics, biologists iteratively mutate antibody sequences to improve binding and stability. Proposed mutations can be informed by previous measurements or by learning from large antibody databases to predict only typical antibodies. Unfortunately, the space of typical antibodies is enormous to search, and experiments often fail to find suitable antibodies on a budget. We introduce Clone-informed Bayesian Optimization (CloneBO), a Bayesian optimization procedure that efficiently optimizes antibodies in the lab by teaching a generative model how our immune system optimizes antibodies. Our immune system makes antibodies by iteratively evolving specific portions of their sequences to bind their target strongly and stably, resulting in a set of related, evolving sequences known as a clonal family. We train a large language model, CloneLM, on hundreds of thousands of clonal families and use it to design sequences with mutations that are most likely to optimize an antibody within the human immune system. We propose to guide our designs to fit previous measurements with a twisted sequential Monte Carlo procedure. We show that CloneBO optimizes antibodies substantially more efficiently than previous methods in realistic in silico experiments and designs stronger and more stable binders in in vitro wet lab experiments.
comment: Code available at https://github.com/AlanNawzadAmin/CloneBO
☆ Efficient Online Reinforcement Learning Fine-Tuning Need Not Retain Offline Data
The modern paradigm in machine learning involves pre-training on diverse data, followed by task-specific fine-tuning. In reinforcement learning (RL), this translates to learning via offline RL on a diverse historical dataset, followed by rapid online RL fine-tuning using interaction data. Most RL fine-tuning methods require continued training on offline data for stability and performance. However, this is undesirable because training on diverse offline data is slow and expensive for large datasets, and in principle, also limit the performance improvement possible because of constraints or pessimism on offline data. In this paper, we show that retaining offline data is unnecessary as long as we use a properly-designed online RL approach for fine-tuning offline RL initializations. To build this approach, we start by analyzing the role of retaining offline data in online fine-tuning. We find that continued training on offline data is mostly useful for preventing a sudden divergence in the value function at the onset of fine-tuning, caused by a distribution mismatch between the offline data and online rollouts. This divergence typically results in unlearning and forgetting the benefits of offline pre-training. Our approach, Warm-start RL (WSRL), mitigates the catastrophic forgetting of pre-trained initializations using a very simple idea. WSRL employs a warmup phase that seeds the online RL run with a very small number of rollouts from the pre-trained policy to do fast online RL. The data collected during warmup helps ``recalibrate'' the offline Q-function to the online distribution, allowing us to completely discard offline data without destabilizing the online RL fine-tuning. We show that WSRL is able to fine-tune without retaining any offline data, and is able to learn faster and attains higher performance than existing algorithms irrespective of whether they retain offline data or not.
☆ PortraitTalk: Towards Customizable One-Shot Audio-to-Talking Face Generation
Audio-driven talking face generation is a challenging task in digital communication. Despite significant progress in the area, most existing methods concentrate on audio-lip synchronization, often overlooking aspects such as visual quality, customization, and generalization that are crucial to producing realistic talking faces. To address these limitations, we introduce a novel, customizable one-shot audio-driven talking face generation framework, named PortraitTalk. Our proposed method utilizes a latent diffusion framework consisting of two main components: IdentityNet and AnimateNet. IdentityNet is designed to preserve identity features consistently across the generated video frames, while AnimateNet aims to enhance temporal coherence and motion consistency. This framework also integrates an audio input with the reference images, thereby reducing the reliance on reference-style videos prevalent in existing approaches. A key innovation of PortraitTalk is the incorporation of text prompts through decoupled cross-attention mechanisms, which significantly expands creative control over the generated videos. Through extensive experiments, including a newly developed evaluation metric, our model demonstrates superior performance over the state-of-the-art methods, setting a new standard for the generation of customizable realistic talking faces suitable for real-world applications.
☆ FlashRNN: Optimizing Traditional RNNs on Modern Hardware
While Transformers and other sequence-parallelizable neural network architectures seem like the current state of the art in sequence modeling, they specifically lack state-tracking capabilities. These are important for time-series tasks and logical reasoning. Traditional RNNs like LSTMs and GRUs, as well as modern variants like sLSTM do have these capabilities at the cost of strictly sequential processing. While this is often seen as a strong limitation, we show how fast these networks can get with our hardware-optimization FlashRNN in Triton and CUDA, optimizing kernels to the register level on modern GPUs. We extend traditional RNNs with a parallelization variant that processes multiple RNNs of smaller hidden state in parallel, similar to the head-wise processing in Transformers. To enable flexibility on different GPU variants, we introduce a new optimization framework for hardware-internal cache sizes, memory and compute handling. It models the hardware in a setting using polyhedral-like constraints, including the notion of divisibility. This speeds up the solution process in our ConstrINT library for general integer constraint satisfaction problems (integer CSPs). We show that our kernels can achieve 50x speed-ups over a vanilla PyTorch implementation and allow 40x larger hidden sizes compared to our Triton implementation. Our open-source kernels and the optimization library are released here to boost research in the direction of state-tracking enabled RNNs and sequence modeling: \url{https://github.com/NX-AI/flashrnn}
☆ Predictive Modeling of Homeless Service Assignment: A Representation Learning Approach
In recent years, there has been growing interest in leveraging machine learning for homeless service assignment. However, the categorical nature of administrative data recorded for homeless individuals hinders the development of accurate machine learning methods for this task. This work asserts that deriving latent representations of such features, while at the same time leveraging underlying relationships between instances is crucial in algorithmically enhancing the existing assignment decision-making process. Our proposed approach learns temporal and functional relationships between services from historical data, as well as unobserved but relevant relationships between individuals to generate features that significantly improve the prediction of the next service assignment compared to the state-of-the-art.
☆ LoRA3D: Low-Rank Self-Calibration of 3D Geometric Foundation Models
Emerging 3D geometric foundation models, such as DUSt3R, offer a promising approach for in-the-wild 3D vision tasks. However, due to the high-dimensional nature of the problem space and scarcity of high-quality 3D data, these pre-trained models still struggle to generalize to many challenging circumstances, such as limited view overlap or low lighting. To address this, we propose LoRA3D, an efficient self-calibration pipeline to $\textit{specialize}$ the pre-trained models to target scenes using their own multi-view predictions. Taking sparse RGB images as input, we leverage robust optimization techniques to refine multi-view predictions and align them into a global coordinate frame. In particular, we incorporate prediction confidence into the geometric optimization process, automatically re-weighting the confidence to better reflect point estimation accuracy. We use the calibrated confidence to generate high-quality pseudo labels for the calibrating views and use low-rank adaptation (LoRA) to fine-tune the models on the pseudo-labeled data. Our method does not require any external priors or manual labels. It completes the self-calibration process on a $\textbf{single standard GPU within just 5 minutes}$. Each low-rank adapter requires only $\textbf{18MB}$ of storage. We evaluated our method on $\textbf{more than 160 scenes}$ from the Replica, TUM and Waymo Open datasets, achieving up to $\textbf{88% performance improvement}$ on 3D reconstruction, multi-view pose estimation and novel-view rendering.
☆ Explainable machine learning for neoplasms diagnosis via electrocardiograms: an externally validated study
Background: Neoplasms remains a leading cause of mortality worldwide, with timely diagnosis being crucial for improving patient outcomes. Current diagnostic methods are often invasive, costly, and inaccessible to many populations. Electrocardiogram (ECG) data, widely available and non-invasive, has the potential to serve as a tool for neoplasms diagnosis by using physiological changes in cardiovascular function associated with neoplastic prescences. Methods: This study explores the application of machine learning models to analyze ECG features for the diagnosis of neoplasms. We developed a pipeline integrating tree-based models with Shapley values for explainability. The model was trained and internally validated and externally validated on a second large-scale independent external cohort to ensure robustness and generalizability. Findings: The results demonstrate that ECG data can effectively capture neoplasms-associated cardiovascular changes, achieving high performance in both internal testing and external validation cohorts. Shapley values identified key ECG features influencing model predictions, revealing established and novel cardiovascular markers linked to neoplastic conditions. This non-invasive approach provides a cost-effective and scalable alternative for the diagnosis of neoplasms, particularly in resource-limited settings. Similarly, useful for the management of secondary cardiovascular effects given neoplasms therapies. Interpretation: This study highlights the feasibility of leveraging ECG signals and machine learning to enhance neoplasms diagnostics. By offering interpretable insights into cardio-neoplasms interactions, this approach bridges existing gaps in non-invasive diagnostics and has implications for integrating ECG-based tools into broader neoplasms diagnostic frameworks, as well as neoplasms therapy management.
comment: 9 pages, 2 figures, code under https://github.com/AI4HealthUOL/CardioDiag
☆ STIV: Scalable Text and Image Conditioned Video Generation
The field of video generation has made remarkable advancements, yet there remains a pressing need for a clear, systematic recipe that can guide the development of robust and scalable models. In this work, we present a comprehensive study that systematically explores the interplay of model architectures, training recipes, and data curation strategies, culminating in a simple and scalable text-image-conditioned video generation method, named STIV. Our framework integrates image condition into a Diffusion Transformer (DiT) through frame replacement, while incorporating text conditioning via a joint image-text conditional classifier-free guidance. This design enables STIV to perform both text-to-video (T2V) and text-image-to-video (TI2V) tasks simultaneously. Additionally, STIV can be easily extended to various applications, such as video prediction, frame interpolation, multi-view generation, and long video generation, etc. With comprehensive ablation studies on T2I, T2V, and TI2V, STIV demonstrate strong performance, despite its simple design. An 8.7B model with 512 resolution achieves 83.1 on VBench T2V, surpassing both leading open and closed-source models like CogVideoX-5B, Pika, Kling, and Gen-3. The same-sized model also achieves a state-of-the-art result of 90.1 on VBench I2V task at 512 resolution. By providing a transparent and extensible recipe for building cutting-edge video generation models, we aim to empower future research and accelerate progress toward more versatile and reliable video generation solutions.
☆ Quantum vs. Classical Machine Learning Algorithms for Software Defect Prediction: Challenges and Opportunities ICSE 2025
Software defect prediction is a critical aspect of software quality assurance, as it enables early identification and mitigation of defects, thereby reducing the cost and impact of software failures. Over the past few years, quantum computing has risen as an exciting technology capable of transforming multiple domains; Quantum Machine Learning (QML) is one of them. QML algorithms harness the power of quantum computing to solve complex problems with better efficiency and effectiveness than their classical counterparts. However, research into its application in software engineering to predict software defects still needs to be explored. In this study, we worked to fill the research gap by comparing the performance of three QML and five classical machine learning (CML) algorithms on the 20 software defect datasets. Our investigation reports the comparative scenarios of QML vs. CML algorithms and identifies the better-performing and consistent algorithms to predict software defects. We also highlight the challenges and future directions of employing QML algorithms in real software defect datasets based on the experience we faced while performing this investigation. The findings of this study can help practitioners and researchers further progress in this research domain by making software systems reliable and bug-free.
comment: In the proceedings of the 6th Quantum Software Engineering (Q-SE) workshop at the 47th IEEE/ACM International Conference on Software Engineering (ICSE 2025)
☆ SimVS: Simulating World Inconsistencies for Robust View Synthesis
Novel-view synthesis techniques achieve impressive results for static scenes but struggle when faced with the inconsistencies inherent to casual capture settings: varying illumination, scene motion, and other unintended effects that are difficult to model explicitly. We present an approach for leveraging generative video models to simulate the inconsistencies in the world that can occur during capture. We use this process, along with existing multi-view datasets, to create synthetic data for training a multi-view harmonization network that is able to reconcile inconsistent observations into a consistent 3D scene. We demonstrate that our world-simulation strategy significantly outperforms traditional augmentation methods in handling real-world scene variations, thereby enabling highly accurate static 3D reconstructions in the presence of a variety of challenging inconsistencies. Project page: https://alextrevithick.github.io/simvs
comment: Project page: https://alextrevithick.github.io/simvs
☆ Privacy-Preserving Customer Support: A Framework for Secure and Scalable Interactions
The growing reliance on artificial intelligence (AI) in customer support has significantly improved operational efficiency and user experience. However, traditional machine learning (ML) approaches, which require extensive local training on sensitive datasets, pose substantial privacy risks and compliance challenges with regulations like the General Data Protection Regulation (GDPR) and California Consumer Privacy Act (CCPA). Existing privacy-preserving techniques, such as anonymization, differential privacy, and federated learning, address some concerns but face limitations in utility, scalability, and complexity. This paper introduces the Privacy-Preserving Zero-Shot Learning (PP-ZSL) framework, a novel approach leveraging large language models (LLMs) in a zero-shot learning mode. Unlike conventional ML methods, PP-ZSL eliminates the need for local training on sensitive data by utilizing pre-trained LLMs to generate responses directly. The framework incorporates real-time data anonymization to redact or mask sensitive information, retrieval-augmented generation (RAG) for domain-specific query resolution, and robust post-processing to ensure compliance with regulatory standards. This combination reduces privacy risks, simplifies compliance, and enhances scalability and operational efficiency. Empirical analysis demonstrates that the PP-ZSL framework provides accurate, privacy-compliant responses while significantly lowering the costs and complexities of deploying AI-driven customer support systems. The study highlights potential applications across industries, including financial services, healthcare, e-commerce, legal support, telecommunications, and government services. By addressing the dual challenges of privacy and performance, this framework establishes a foundation for secure, efficient, and regulatory-compliant AI applications in customer interactions.
☆ Optimizing Sensor Redundancy in Sequential Decision-Making Problems
Reinforcement Learning (RL) policies are designed to predict actions based on current observations to maximize cumulative future rewards. In real-world applications (i.e., non-simulated environments), sensors are essential for measuring the current state and providing the observations on which RL policies rely to make decisions. A significant challenge in deploying RL policies in real-world scenarios is handling sensor dropouts, which can result from hardware malfunctions, physical damage, or environmental factors like dust on a camera lens. A common strategy to mitigate this issue is the use of backup sensors, though this comes with added costs. This paper explores the optimization of backup sensor configurations to maximize expected returns while keeping costs below a specified threshold, C. Our approach uses a second-order approximation of expected returns and includes penalties for exceeding cost constraints. We then optimize this quadratic program using Tabu Search, a meta-heuristic algorithm. The approach is evaluated across eight OpenAI Gym environments and a custom Unity-based robotic environment (RobotArmGrasping). Empirical results demonstrate that our quadratic program effectively approximates real expected returns, facilitating the identification of optimal sensor configurations.
comment: Accepted at ICAART conference 2025
☆ The Pitfalls of Memorization: When Memorization Hurts Generalization
Neural networks often learn simple explanations that fit the majority of the data while memorizing exceptions that deviate from these explanations.This behavior leads to poor generalization when the learned explanations rely on spurious correlations. In this work, we formalize the interplay between memorization and generalization, showing that spurious correlations would particularly lead to poor generalization when are combined with memorization. Memorization can reduce training loss to zero, leaving no incentive to learn robust, generalizable patterns. To address this, we propose memorization-aware training (MAT), which uses held-out predictions as a signal of memorization to shift a model's logits. MAT encourages learning robust patterns invariant across distributions, improving generalization under distribution shifts.
☆ RAZOR: Sharpening Knowledge by Cutting Bias with Unsupervised Text Rewriting AAAI'25
Despite the widespread use of LLMs due to their superior performance in various tasks, their high computational costs often lead potential users to opt for the pretraining-finetuning pipeline. However, biases prevalent in manually constructed datasets can introduce spurious correlations between tokens and labels, creating so-called shortcuts and hindering the generalizability of fine-tuned models. Existing debiasing methods often rely on prior knowledge of specific dataset biases, which is challenging to acquire a priori. We propose RAZOR (Rewriting And Zero-bias Optimization Refinement), a novel, unsupervised, and data-focused debiasing approach based on text rewriting for shortcut mitigation. RAZOR leverages LLMs to iteratively rewrite potentially biased text segments by replacing them with heuristically selected alternatives in a shortcut space defined by token statistics and positional information. This process aims to align surface-level text features more closely with diverse label distributions, thereby promoting the learning of genuine linguistic patterns. Compared with unsupervised SoTA models, RAZOR improves by 3.5% on the FEVER and 6.5% on MNLI and SNLI datasets according to the F1 score. Additionally, RAZOR effectively mitigates specific known biases, reducing bias-related terms by x2 without requiring prior bias information, a result that is on par with SoTA models that leverage prior information. Our work prioritizes data manipulation over architectural modifications, emphasizing the pivotal role of data quality in enhancing model performance and fairness. This research contributes to developing more robust evaluation benchmarks for debiasing methods by incorporating metrics for bias reduction and overall model efficacy.
comment: Shuo and Bardh contributed equally. Accepted to AAAI'25
☆ TraSCE: Trajectory Steering for Concept Erasure
Recent advancements in text-to-image diffusion models have brought them to the public spotlight, becoming widely accessible and embraced by everyday users. However, these models have been shown to generate harmful content such as not-safe-for-work (NSFW) images. While approaches have been proposed to erase such abstract concepts from the models, jail-breaking techniques have succeeded in bypassing such safety measures. In this paper, we propose TraSCE, an approach to guide the diffusion trajectory away from generating harmful content. Our approach is based on negative prompting, but as we show in this paper, conventional negative prompting is not a complete solution and can easily be bypassed in some corner cases. To address this issue, we first propose a modification of conventional negative prompting. Furthermore, we introduce a localized loss-based guidance that enhances the modified negative prompting technique by steering the diffusion trajectory. We demonstrate that our proposed method achieves state-of-the-art results on various benchmarks in removing harmful content including ones proposed by red teams; and erasing artistic styles and objects. Our proposed approach does not require any training, weight modifications, or training data (both image or prompt), making it easier for model owners to erase new concepts.
☆ Bayesian Data Augmentation and Training for Perception DNN in Autonomous Aerial Vehicles
Learning-based solutions have enabled incredible capabilities for autonomous systems. Autonomous vehicles, both aerial and ground, rely on DNN for various integral tasks, including perception. The efficacy of supervised learning solutions hinges on the quality of the training data. Discrepancies between training data and operating conditions result in faults that can lead to catastrophic incidents. However, collecting vast amounts of context-sensitive data, with broad coverage of possible operating environments, is prohibitively difficult. Synthetic data generation techniques for DNN allow for the easy exploration of diverse scenarios. However, synthetic data generation solutions for aerial vehicles are still lacking. This work presents a data augmentation framework for aerial vehicle's perception training, leveraging photorealistic simulation integrated with high-fidelity vehicle dynamics. Safe landing is a crucial challenge in the development of autonomous air taxis, therefore, landing maneuver is chosen as the focus of this work. With repeated simulations of landing in varying scenarios we assess the landing performance of the VTOL type UAV and gather valuable data. The landing performance is used as the objective function to optimize the DNN through retraining. Given the high computational cost of DNN retraining, we incorporated Bayesian Optimization in our framework that systematically explores the data augmentation parameter space to retrain the best-performing models. The framework allowed us to identify high-performing data augmentation parameters that are consistently effective across different landing scenarios. Utilizing the capabilities of this data augmentation framework, we obtained a robust perception model. The model consistently improved the perception-based landing success rate by at least 20% under different lighting and weather conditions.
comment: To be published in AIAA SciTech 2025 Forum
☆ Offline Multi-Agent Reinforcement Learning via In-Sample Sequential Policy Optimization
Offline Multi-Agent Reinforcement Learning (MARL) is an emerging field that aims to learn optimal multi-agent policies from pre-collected datasets. Compared to single-agent case, multi-agent setting involves a large joint state-action space and coupled behaviors of multiple agents, which bring extra complexity to offline policy optimization. In this work, we revisit the existing offline MARL methods and show that in certain scenarios they can be problematic, leading to uncoordinated behaviors and out-of-distribution (OOD) joint actions. To address these issues, we propose a new offline MARL algorithm, named In-Sample Sequential Policy Optimization (InSPO). InSPO sequentially updates each agent's policy in an in-sample manner, which not only avoids selecting OOD joint actions but also carefully considers teammates' updated policies to enhance coordination. Additionally, by thoroughly exploring low-probability actions in the behavior policy, InSPO can well address the issue of premature convergence to sub-optimal solutions. Theoretically, we prove InSPO guarantees monotonic policy improvement and converges to quantal response equilibrium (QRE). Experimental results demonstrate the effectiveness of our method compared to current state-of-the-art offline MARL methods.
☆ SurvBETA: Ensemble-Based Survival Models Using Beran Estimators and Several Attention Mechanisms
Many ensemble-based models have been proposed to solve machine learning problems in the survival analysis framework, including random survival forests, the gradient boosting machine with weak survival models, ensembles of the Cox models. To extend the set of models, a new ensemble-based model called SurvBETA (the Survival Beran estimator Ensemble using Three Attention mechanisms) is proposed where the Beran estimator is used as a weak learner in the ensemble. The Beran estimator can be regarded as a kernel regression model taking into account the relationship between instances. Outputs of weak learners in the form of conditional survival functions are aggregated with attention weights taking into account the distance between the analyzed instance and prototypes of all bootstrap samples. The attention mechanism is used three times: for implementation of the Beran estimators, for determining specific prototypes of bootstrap samples and for aggregating the weak model predictions. The proposed model is presented in two forms: in a general form requiring to solve a complex optimization problem for its training; in a simplified form by considering a special representation of the attention weights by means of the imprecise Huber's contamination model which leads to solving a simple optimization problem. Numerical experiments illustrate properties of the model on synthetic data and compare the model with other survival models on real data. A code implementing the proposed model is publicly available.
☆ Sampling from Boltzmann densities with physics informed low-rank formats
Our method proposes the efficient generation of samples from an unnormalized Boltzmann density by solving the underlying continuity equation in the low-rank tensor train (TT) format. It is based on the annealing path commonly used in MCMC literature, which is given by the linear interpolation in the space of energies. Inspired by Sequential Monte Carlo, we alternate between deterministic time steps from the TT representation of the flow field and stochastic steps, which include Langevin and resampling steps. These adjust the relative weights of the different modes of the target distribution and anneal to the correct path distribution. We showcase the efficiency of our method on multiple numerical examples.
☆ Fast Track to Winning Tickets: Repowering One-Shot Pruning for Graph Neural Networks AAAI 2025
Graph Neural Networks (GNNs) demonstrate superior performance in various graph learning tasks, yet their wider real-world application is hindered by the computational overhead when applied to large-scale graphs. To address the issue, the Graph Lottery Hypothesis (GLT) has been proposed, advocating the identification of subgraphs and subnetworks, \textit{i.e.}, winning tickets, without compromising performance. The effectiveness of current GLT methods largely stems from the use of iterative magnitude pruning (IMP), which offers higher stability and better performance than one-shot pruning. However, identifying GLTs is highly computationally expensive, due to the iterative pruning and retraining required by IMP. In this paper, we reevaluate the correlation between one-shot pruning and IMP: while one-shot tickets are suboptimal compared to IMP, they offer a \textit{fast track} to tickets with a stronger performance. We introduce a one-shot pruning and denoising framework to validate the efficacy of the \textit{fast track}. Compared to current IMP-based GLT methods, our framework achieves a double-win situation of graph lottery tickets with \textbf{higher sparsity} and \textbf{faster speeds}. Through extensive experiments across 4 backbones and 6 datasets, our method demonstrates $1.32\% - 45.62\%$ improvement in weight sparsity and a $7.49\% - 22.71\%$ increase in graph sparsity, along with a $1.7-44 \times$ speedup over IMP-based methods and $95.3\%-98.6\%$ MAC savings.
comment: AAAI 2025
☆ Hype-Adjusted Probability Measure for NLP Volatility Forecasting
This manuscript introduces the hype-adjusted probability measure developed in the context of a new Natural Language Processing (NLP) approach for market forecasting. A novel sentiment score equation is presented to capture component and memory effects and assign dynamic parameters, enhancing the impact of intraday news data on forecasting next-period volatility for selected U.S. semiconductor stocks. This approach integrates machine learning techniques to analyze and improve the predictive value of news. Building on the research of Geman's, this work improves forecast accuracy by assigning specific weights to each component of news sources and individual stocks in the portfolio, evaluating time-memory effects on market reactions, and incorporating shifts in sentiment direction. Finally, we propose the Hype-Adjusted Probability Measure, proving its existence and uniqueness, and discuss its theoretical applications in finance for NLP-based volatility forecasting, outlining future research pathways inspired by its concepts.
comment: 26 pages
☆ Paired Wasserstein Autoencoders for Conditional Sampling
Wasserstein distances greatly influenced and coined various types of generative neural network models. Wasserstein autoencoders are particularly notable for their mathematical simplicity and straight-forward implementation. However, their adaptation to the conditional case displays theoretical difficulties. As a remedy, we propose the use of two paired autoencoders. Under the assumption of an optimal autoencoder pair, we leverage the pairwise independence condition of our prescribed Gaussian latent distribution to overcome this theoretical hurdle. We conduct several experiments to showcase the practical applicability of the resulting paired Wasserstein autoencoders. Here, we consider imaging tasks and enable conditional sampling for denoising, inpainting, and unsupervised image translation. Moreover, we connect our image translation model to the Monge map behind Wasserstein-2 distances.
☆ Scaling Sequential Recommendation Models with Transformers
Modeling user preferences has been mainly addressed by looking at users' interaction history with the different elements available in the system. Tailoring content to individual preferences based on historical data is the main goal of sequential recommendation. The nature of the problem, as well as the good performance observed across various domains, has motivated the use of the transformer architecture, which has proven effective in leveraging increasingly larger amounts of training data when accompanied by an increase in the number of model parameters. This scaling behavior has brought a great deal of attention, as it provides valuable guidance in the design and training of even larger models. Taking inspiration from the scaling laws observed in training large language models, we explore similar principles for sequential recommendation. We use the full Amazon Product Data dataset, which has only been partially explored in other studies, and reveal scaling behaviors similar to those found in language models. Compute-optimal training is possible but requires a careful analysis of the compute-performance trade-offs specific to the application. We also show that performance scaling translates to downstream tasks by fine-tuning larger pre-trained models on smaller task-specific domains. Our approach and findings provide a strategic roadmap for model training and deployment in real high-dimensional preference spaces, facilitating better training and inference efficiency. We hope this paper bridges the gap between the potential of transformers and the intrinsic complexities of high-dimensional sequential recommendation in real-world recommender systems. Code and models can be found at https://github.com/mercadolibre/srt
☆ Adaptive Epsilon Adversarial Training for Robust Gravitational Wave Parameter Estimation Using Normalizing Flows
Adversarial training with Normalizing Flow (NF) models is an emerging research area aimed at improving model robustness through adversarial samples. In this study, we focus on applying adversarial training to NF models for gravitational wave parameter estimation. We propose an adaptive epsilon method for Fast Gradient Sign Method (FGSM) adversarial training, which dynamically adjusts perturbation strengths based on gradient magnitudes using logarithmic scaling. Our hybrid architecture, combining ResNet and Inverse Autoregressive Flow, reduces the Negative Log Likelihood (NLL) loss by 47\% under FGSM attacks compared to the baseline model, while maintaining an NLL of 4.2 on clean data (only 5\% higher than the baseline). For perturbation strengths between 0.01 and 0.1, our model achieves an average NLL of 5.8, outperforming both fixed-epsilon (NLL: 6.7) and progressive-epsilon (NLL: 7.2) methods. Under stronger Projected Gradient Descent attacks with perturbation strength of 0.05, our model maintains an NLL of 6.4, demonstrating superior robustness while avoiding catastrophic overfitting.
comment: 7 pages, 9 figures
☆ Contractive Dynamical Imitation Policies for Efficient Out-of-Sample Recovery
Imitation learning is a data-driven approach to learning policies from expert behavior, but it is prone to unreliable outcomes in out-of-sample (OOS) regions. While previous research relying on stable dynamical systems guarantees convergence to a desired state, it often overlooks transient behavior. We propose a framework for learning policies using modeled by contractive dynamical systems, ensuring that all policy rollouts converge regardless of perturbations, and in turn, enable efficient OOS recovery. By leveraging recurrent equilibrium networks and coupling layers, the policy structure guarantees contractivity for any parameter choice, which facilitates unconstrained optimization. Furthermore, we provide theoretical upper bounds for worst-case and expected loss terms, rigorously establishing the reliability of our method in deployment. Empirically, we demonstrate substantial OOS performance improvements in robotics manipulation and navigation tasks in simulation.
☆ Anomaly detection using Diffusion-based methods
This paper explores the utility of diffusion-based models for anomaly detection, focusing on their efficacy in identifying deviations in both compact and high-resolution datasets. Diffusion-based architectures, including Denoising Diffusion Probabilistic Models (DDPMs) and Diffusion Transformers (DiTs), are evaluated for their performance using reconstruction objectives. By leveraging the strengths of these models, this study benchmarks their performance against traditional anomaly detection methods such as Isolation Forests, One-Class SVMs, and COPOD. The results demonstrate the superior adaptability, scalability, and robustness of diffusion-based methods in handling complex real-world anomaly detection tasks. Key findings highlight the role of reconstruction error in enhancing detection accuracy and underscore the scalability of these models to high-dimensional datasets. Future directions include optimizing encoder-decoder architectures and exploring multi-modal datasets to further advance diffusion-based anomaly detection.
☆ Can Neural Decompilation Assist Vulnerability Prediction on Binary Code?
Vulnerability prediction is valuable in identifying security issues more efficiently, even though it requires the source code of the target software system, which is a restrictive hypothesis. This paper presents an experimental study to predict vulnerabilities in binary code without source code or complex representations of the binary, leveraging the pivotal idea of decompiling the binary file through neural decompilation and predicting vulnerabilities through deep learning on the decompiled source code. The results outperform the state-of-the-art in both neural decompilation and vulnerability prediction, showing that it is possible to identify vulnerable programs with this approach concerning bi-class (vulnerable/non-vulnerable) and multi-class (type of vulnerability) analysis.
☆ Quantifying the Prediction Uncertainty of Machine Learning Models for Individual Data
Machine learning models have exhibited exceptional results in various domains. The most prevalent approach for learning is the empirical risk minimizer (ERM), which adapts the model's weights to reduce the loss on a training set and subsequently leverages these weights to predict the label for new test data. Nonetheless, ERM makes the assumption that the test distribution is similar to the training distribution, which may not always hold in real-world situations. In contrast, the predictive normalized maximum likelihood (pNML) was proposed as a min-max solution for the individual setting where no assumptions are made on the distribution of the tested input. This study investigates pNML's learnability for linear regression and neural networks, and demonstrates that pNML can improve the performance and robustness of these models on various tasks. Moreover, the pNML provides an accurate confidence measure for its output, showcasing state-of-the-art results for out-of-distribution detection, resistance to adversarial attacks, and active learning.
comment: PHD thesis
☆ Physics-Based Dynamic Models Hybridisation Using Physics-Informed Neural Networks
Physics-based dynamic models (PBDMs) are simplified representations of complex dynamical systems. PBDMs take specific processes within a complex system and assign a fragment of variables and an accompanying set of parameters to depict the processes. As this often leads to suboptimal parameterisation of the system, a key challenge requires refining the empirical parameters and variables to reduce uncertainties while maintaining the model s explainability and enhancing its predictive accuracy. We demonstrate that a hybrid mosquito population dynamics model, which integrates a PBDM with Physics-Informed Neural Networks (PINN), retains the explainability of the PBDM by incorporating the PINN-learned model parameters in place of its empirical counterparts. Specifically, we address the limitations of traditional PBDMs by modelling the parameters of larva and pupa development rates using a PINN that encodes complex, learned interactions of air temperature, precipitation and humidity. Our results demonstrate improved mosquito population simulations including the difficult-to-predict mosquito population peaks. This opens the possibility of hybridisation concept application on other complex systems based on PBDMs such as cancer growth to address the challenges posed by scarce and noisy data, and to numerical weather prediction and climate modelling to overcome the gap between physics-based and data-driven weather prediction models.
☆ ConfigX: Modular Configuration for Evolutionary Algorithms via Multitask Reinforcement Learning
Recent advances in Meta-learning for Black-Box Optimization (MetaBBO) have shown the potential of using neural networks to dynamically configure evolutionary algorithms (EAs), enhancing their performance and adaptability across various BBO instances. However, they are often tailored to a specific EA, which limits their generalizability and necessitates retraining or redesigns for different EAs and optimization problems. To address this limitation, we introduce ConfigX, a new paradigm of the MetaBBO framework that is capable of learning a universal configuration agent (model) for boosting diverse EAs. To achieve so, our ConfigX first leverages a novel modularization system that enables the flexible combination of various optimization sub-modules to generate diverse EAs during training. Additionally, we propose a Transformer-based neural network to meta-learn a universal configuration policy through multitask reinforcement learning across a designed joint optimization task space. Extensive experiments verify that, our ConfigX, after large-scale pre-training, achieves robust zero-shot generalization to unseen tasks and outperforms state-of-the-art baselines. Moreover, ConfigX exhibits strong lifelong learning capabilities, allowing efficient adaptation to new tasks through fine-tuning. Our proposed ConfigX represents a significant step toward an automatic, all-purpose configuration agent for EAs.
☆ Dual Random Fields and their Application to Mineral Potential Mapping
In various geosciences branches, including mineral exploration, geometallurgical characterization on established mining operations, and remote sensing, the regionalized input variables are spatially well-sampled across the domain of interest, limiting the scope of spatial uncertainty quantification procedures. In turn, response outcomes such as the mineral potential in a given region, mining throughput, metallurgical recovery, or in-situ estimations from remote satellite imagery, are usually modeled from a much-restricted subset of testing samples, collected at certain locations due to accessibility restrictions and the high acquisition costs. Our limited understanding of these functions, in terms of the multi-dimensional complexity of causalities and unnoticed dependencies on inaccessible inputs, may lead to observing changes in such functions based on their geographical location. Pooling together different response functions across the domain is critical to correctly predict outcome responses, the uncertainty associated with these inferred values, and the significance of inputs in such predictions at unexplored areas. This paper introduces the notion of a dual random field (dRF), where the response function itself is considered a regionalized variable. In this way, different established response models across the geographic domain can be considered as observations of a dRF realization, enabling the spatial inference and uncertainty assessment of both response models and their predictions. We explain how dRFs inherit all the properties from classical random fields, allowing the use of standard Gaussian simulation procedures to simulate them. These models are combined to obtain a mineral potential response, providing an example of how to rigorously integrate machine learning approaches with geostatistics.
☆ Real-time Sign Language Recognition Using MobileNetV2 and Transfer Learning
The hearing-impaired community in India deserves the access to tools that help them communicate, however, there is limited known technology solutions that make use of Indian Sign Language (ISL) at present. Even though there are many ISL users, ISL cannot access social and education arenas because there is not yet an efficient technology to convert the ISL signal into speech or text. We initiated this initiative owing to the rising demand for products and technologies that are inclusive and help ISL, filling the gap of communication for the ones with hearing disability. Our goal is to build an reliable sign language recognition system with the help of Convolutional Neural Networks (CNN) to . By expanding communication access, we aspire toward better educational opportunities and a more inclusive society for hearing impaired people in India.
☆ Progressive-Resolution Policy Distillation: Leveraging Coarse-Resolution Simulation for Time-Efficient Fine-Resolution Policy Learning
In earthwork and construction, excavators often encounter large rocks mixed with various soil conditions, requiring skilled operators. This paper presents a framework for achieving autonomous excavation using reinforcement learning (RL) through a rock excavation simulator. In the simulation, resolution can be defined by the particle size/number in the whole soil space. Fine-resolution simulations closely mimic real-world behavior but demand significant calculation time and challenging sample collection, while coarse-resolution simulations enable faster sample collection but deviate from real-world behavior. To combine the advantages of both resolutions, we explore using policies developed in coarse-resolution simulations for pre-training in fine-resolution simulations. To this end, we propose a novel policy learning framework called Progressive-Resolution Policy Distillation (PRPD), which progressively transfers policies through some middle-resolution simulations with conservative policy transfer to avoid domain gaps that could lead to policy transfer failure. Validation in a rock excavation simulator and nine real-world rock environments demonstrated that PRPD reduced sampling time to less than 1/7 while maintaining task success rates comparable to those achieved through policy learning in a fine-resolution simulation.
☆ Score-matching-based Structure Learning for Temporal Data on Networks
Causal discovery is a crucial initial step in establishing causality from empirical data and background knowledge. Numerous algorithms have been developed for this purpose. Among them, the score-matching method has demonstrated superior performance across various evaluation metrics, particularly for the commonly encountered Additive Nonlinear Causal Models. However, current score-matching-based algorithms are primarily designed to analyze independent and identically distributed (i.i.d.) data. More importantly, they suffer from high computational complexity due to the pruning step required for handling dense Directed Acyclic Graphs (DAGs). To enhance the scalability of score matching, we have developed a new parent-finding subroutine for leaf nodes in DAGs, significantly accelerating the most time-consuming part of the process: the pruning step. This improvement results in an efficiency-lifted score matching algorithm, termed Parent Identification-based Causal structure learning for both i.i.d. and temporal data on networKs, or PICK. The new score-matching algorithm extends the scope of existing algorithms and can handle static and temporal data on networks with weak network interference. Our proposed algorithm can efficiently cope with increasingly complex datasets that exhibit spatial and temporal dependencies, commonly encountered in academia and industry. The proposed algorithm can accelerate score-matching-based methods while maintaining high accuracy in real-world applications.
☆ AHSG: Adversarial Attacks on High-level Semantics in Graph Neural Networks
Graph Neural Networks (GNNs) have garnered significant interest among researchers due to their impressive performance in graph learning tasks. However, like other deep neural networks, GNNs are also vulnerable to adversarial attacks. In existing adversarial attack methods for GNNs, the metric between the attacked graph and the original graph is usually the attack budget or a measure of global graph properties. However, we have found that it is possible to generate attack graphs that disrupt the primary semantics even within these constraints. To address this problem, we propose a Adversarial Attacks on High-level Semantics in Graph Neural Networks (AHSG), which is a graph structure attack model that ensures the retention of primary semantics. The latent representations of each node can extract rich semantic information by applying convolutional operations on graph data. These representations contain both task-relevant primary semantic information and task-irrelevant secondary semantic information. The latent representations of same-class nodes with the same primary semantics can fulfill the objective of modifying secondary semantics while preserving the primary semantics. Finally, the latent representations with attack effects is mapped to an attack graph using Projected Gradient Descent (PGD) algorithm. By attacking graph deep learning models with some advanced defense strategies, we validate that AHSG has superior attack effectiveness compared to other attack methods. Additionally, we employ Contextual Stochastic Block Models (CSBMs) as a proxy for the primary semantics to detect the attacked graph, confirming that AHSG almost does not disrupt the original primary semantics of the graph.
☆ Tazza: Shuffling Neural Network Parameters for Secure and Private Federated Learning
Federated learning enables decentralized model training without sharing raw data, preserving data privacy. However, its vulnerability towards critical security threats, such as gradient inversion and model poisoning by malicious clients, remain unresolved. Existing solutions often address these issues separately, sacrificing either system robustness or model accuracy. This work introduces Tazza, a secure and efficient federated learning framework that simultaneously addresses both challenges. By leveraging the permutation equivariance and invariance properties of neural networks via weight shuffling and shuffled model validation, Tazza enhances resilience against diverse poisoning attacks, while ensuring data confidentiality and high model accuracy. Comprehensive evaluations on various datasets and embedded platforms show that Tazza achieves robust defense with up to 6.7x improved computational efficiency compared to alternative schemes, without compromising performance.
comment: 14 pages, 14 figures
☆ Causal World Representation in the GPT Model NeurIPS 2024
Are generative pre-trained transformer (GPT) models only trained to predict the next token, or do they implicitly learn a world model from which a sequence is generated one token at a time? We examine this question by deriving a causal interpretation of the attention mechanism in GPT, and suggesting a causal world model that arises from this interpretation. Furthermore, we propose that GPT-models, at inference time, can be utilized for zero-shot causal structure learning for in-distribution sequences. Empirical evaluation is conducted in a controlled synthetic environment using the setup and rules of the Othello board game. A GPT, pre-trained on real-world games played with the intention of winning, is tested on synthetic data that only adheres to the game rules. We find that the GPT model tends to generate next moves that adhere to the game rules for sequences for which the attention mechanism encodes a causal structure with high confidence. In general, in cases for which the GPT model generates moves that do not adhere to the game rules, it also fails to capture any causal structure.
comment: NeurIPS 2024 Workshop on Causality and Large Models (CaLM)
☆ Reconstructing Deep Neural Networks: Unleashing the Optimization Potential of Natural Gradient Descent
Natural gradient descent (NGD) is a powerful optimization technique for machine learning, but the computational complexity of the inverse Fisher information matrix limits its application in training deep neural networks. To overcome this challenge, we propose a novel optimization method for training deep neural networks called structured natural gradient descent (SNGD). Theoretically, we demonstrate that optimizing the original network using NGD is equivalent to using fast gradient descent (GD) to optimize the reconstructed network with a structural transformation of the parameter matrix. Thereby, we decompose the calculation of the global Fisher information matrix into the efficient computation of local Fisher matrices via constructing local Fisher layers in the reconstructed network to speed up the training. Experimental results on various deep networks and datasets demonstrate that SNGD achieves faster convergence speed than NGD while retaining comparable solutions. Furthermore, our method outperforms traditional GDs in terms of efficiency and effectiveness. Thus, our proposed method has the potential to significantly improve the scalability and efficiency of NGD in deep learning applications. Our source code is available at https://github.com/Chaochao-Lin/SNGD.
☆ Impact of Sampling Techniques and Data Leakage on XGBoost Performance in Credit Card Fraud Detection
Credit card fraud detection remains a critical challenge in financial security, with machine learning models like XGBoost(eXtreme gradient boosting) emerging as powerful tools for identifying fraudulent transactions. However, the inherent class imbalance in credit card transaction datasets poses significant challenges for model performance. Although sampling techniques are commonly used to address this imbalance, their implementation sometimes precedes the train-test split, potentially introducing data leakage. This study presents a comparative analysis of XGBoost's performance in credit card fraud detection under three scenarios: Firstly without any imbalance handling techniques, secondly with sampling techniques applied only to the training set after the train-test split, and third with sampling techniques applied before the train-test split. We utilized a dataset from Kaggle of 284,807 credit card transactions, containing 0.172\% fraudulent cases, to evaluate these approaches. Our findings show that although sampling strategies enhance model performance, the reliability of results is greatly impacted by when they are applied. Due to a data leakage issue that frequently occurs in machine learning models during the sampling phase, XGBoost models trained on data where sampling was applied prior to the train-test split may have displayed artificially inflated performance metrics. Surprisingly, models trained with sampling techniques applied solely to the training set demonstrated significantly lower results than those with pre-split sampling, all the while preserving the integrity of the evaluation process.
comment: 19 pages, 4 figures
☆ Parallel simulation for sampling under isoperimetry and score-based diffusion models
In recent years, there has been a surge of interest in proving discretization bounds for sampling under isoperimetry and for diffusion models. As data size grows, reducing the iteration cost becomes an important goal. Inspired by the great success of the parallel simulation of the initial value problem in scientific computation, we propose parallel Picard methods for sampling tasks. Rigorous theoretical analysis reveals that our algorithm achieves better dependence on dimension $d$ than prior works in iteration complexity (i.e., reduced from $\widetilde{O}(\log^2 d)$ to $\widetilde{O}(\log d)$), which is even optimal for sampling under isoperimetry with specific iteration complexity. Our work highlights the potential advantages of simulation methods in scientific computation for dynamics-based sampling and diffusion models.
comment: arXiv admin note: substantial text overlap with arXiv:2405.15986 by other authors
☆ When UAV Meets Federated Learning: Latency Minimization via Joint Trajectory Design and Resource Allocation
Federated learning (FL) has emerged as a pivotal solution for training machine learning models over wireless networks, particularly for Internet of Things (IoT) devices with limited computation resources. Despite its benefits, the efficiency of FL is often restricted by the communication quality between IoT devices and the central server. To address this issue, we introduce an innovative approach by deploying an unmanned aerial vehicle (UAV) as a mobile FL server to enhance the training process of FL. By leveraging the UAV's maneuverability, we establish robust line-of-sight connections with IoT devices, significantly improving communication capacity. To improve the overall training efficiency, we formulate a latency minimization problem by jointly optimizing the bandwidth allocation, computing frequencies, transmit power for both the UAV and IoT devices, and the UAV's trajectory. Then, an efficient alternating optimization algorithm is developed to solve it efficiently. Furthermore, we analyze the convergence and computational complexity of the proposed algorithm. Finally, numerical results demonstrate that our proposed scheme not only outperforms existing benchmark schemes in terms of latency but also achieves training efficiency that closely approximate the ideal scenario.
comment: This manuscript has been submitted to IEEE
☆ Machine Learning Algorithms for Detecting Mental Stress in College Students
In today's world, stress is a big problem that affects people's health and happiness. More and more people are feeling stressed out, which can lead to lots of health issues like breathing problems, feeling overwhelmed, heart attack, diabetes, etc. This work endeavors to forecast stress and non-stress occurrences among college students by applying various machine learning algorithms: Decision Trees, Random Forest, Support Vector Machines, AdaBoost, Naive Bayes, Logistic Regression, and K-nearest Neighbors. The primary objective of this work is to leverage a research study to predict and mitigate stress and non-stress based on the collected questionnaire dataset. We conducted a workshop with the primary goal of studying the stress levels found among the students. This workshop was attended by Approximately 843 students aged between 18 to 21 years old. A questionnaire was given to the students validated under the guidance of the experts from the All India Institute of Medical Sciences (AIIMS) Raipur, Chhattisgarh, India, on which our dataset is based. The survey consists of 28 questions, aiming to comprehensively understand the multidimensional aspects of stress, including emotional well-being, physical health, academic performance, relationships, and leisure. This work finds that Support Vector Machines have a maximum accuracy for Stress, reaching 95\%. The study contributes to a deeper understanding of stress determinants. It aims to improve college student's overall quality of life and academic success, addressing the multifaceted nature of stress.
comment: This paper was presented at an IEEE conference and is 5 pages long with 5 figures. It discusses machine learning algorithms for detecting mental stress in college students
☆ DSFEC: Efficient and Deployable Deep Radar Object Detection
Deploying radar object detection models on resource-constrained edge devices like the Raspberry Pi poses significant challenges due to the large size of the model and the limited computational power and the memory of the Pi. In this work, we explore the efficiency of Depthwise Separable Convolutions in radar object detection networks and integrate them into our model. Additionally, we introduce a novel Feature Enhancement and Compression (FEC) module to the PointPillars feature encoder to further improve the model performance. With these innovations, we propose the DSFEC-L model and its two versions, which outperform the baseline (23.9 mAP of Car class, 20.72 GFLOPs) on nuScenes dataset: 1). An efficient DSFEC-M model with a 14.6% performance improvement and a 60% reduction in GFLOPs. 2). A deployable DSFEC-S model with a 3.76% performance improvement and a remarkable 78.5% reduction in GFLOPs. Despite marginal performance gains, our deployable model achieves an impressive 74.5% reduction in runtime on the Raspberry Pi compared to the baseline.
☆ Towards Graph Foundation Models: A Study on the Generalization of Positional and Structural Encodings
Recent advances in integrating positional and structural encodings (PSEs) into graph neural networks (GNNs) have significantly enhanced their performance across various graph learning tasks. However, the general applicability of these encodings and their potential to serve as foundational representations for graphs remain uncertain. This paper investigates the fine-tuning efficiency, scalability with sample size, and generalization capability of learnable PSEs across diverse graph datasets. Specifically, we evaluate their potential as universal pre-trained models that can be easily adapted to new tasks with minimal fine-tuning and limited data. Furthermore, we assess the expressivity of the learned representations, particularly, when used to augment downstream GNNs. We demonstrate through extensive benchmarking and empirical analysis that PSEs generally enhance downstream models. However, some datasets may require specific PSE-augmentations to achieve optimal performance. Nevertheless, our findings highlight their significant potential to become integral components of future graph foundation models. We provide new insights into the strengths and limitations of PSEs, contributing to the broader discourse on foundation models in graph learning.
☆ MoDULA: Mixture of Domain-Specific and Universal LoRA for Multi-Task Learning
The growing demand for larger-scale models in the development of \textbf{L}arge \textbf{L}anguage \textbf{M}odels (LLMs) poses challenges for efficient training within limited computational resources. Traditional fine-tuning methods often exhibit instability in multi-task learning and rely heavily on extensive training resources. Here, we propose MoDULA (\textbf{M}ixture \textbf{o}f \textbf{D}omain-Specific and \textbf{U}niversal \textbf{L}oR\textbf{A}), a novel \textbf{P}arameter \textbf{E}fficient \textbf{F}ine-\textbf{T}uning (PEFT) \textbf{M}ixture-\textbf{o}f-\textbf{E}xpert (MoE) paradigm for improved fine-tuning and parameter efficiency in multi-task learning. The paradigm effectively improves the multi-task capability of the model by training universal experts, domain-specific experts, and routers separately. MoDULA-Res is a new method within the MoDULA paradigm, which maintains the model's general capability by connecting universal and task-specific experts through residual connections. The experimental results demonstrate that the overall performance of the MoDULA-Flan and MoDULA-Res methods surpasses that of existing fine-tuning methods on various LLMs. Notably, MoDULA-Res achieves more significant performance improvements in multiple tasks while reducing training costs by over 80\% without losing general capability. Moreover, MoDULA displays flexible pluggability, allowing for the efficient addition of new tasks without retraining existing experts from scratch. This progressive training paradigm circumvents data balancing issues, enhancing training efficiency and model stability. Overall, MoDULA provides a scalable, cost-effective solution for fine-tuning LLMs with enhanced parameter efficiency and generalization capability.
☆ Post-Training Non-Uniform Quantization for Convolutional Neural Networks
Despite the success of CNN models on a variety of Image classification and segmentation tasks, their extensive computational and storage demands pose considerable challenges for real-world deployment on resource constrained devices. Quantization is one technique that aims to alleviate these large storage requirements and speed up the inference process by reducing the precision of model parameters to lower-bit representations. In this paper, we introduce a novel post-training quantization method for model weights. Our method finds optimal clipping thresholds and scaling factors along with mathematical guarantees that our method minimizes quantization noise. Empirical results on Real World Datasets demonstrate that our quantization scheme significantly reduces model size and computational requirements while preserving model accuracy.
☆ Label up: Learning Pulmonary Embolism Segmentation from Image Level Annotation through Model Explainability
Pulmonary Embolisms (PE) are a leading cause of cardiovascular death. Computed tomographic pulmonary angiography (CTPA) stands as the gold standard for diagnosing pulmonary embolisms (PE) and there has been a lot of interest in developing AI-based models for assisting in PE diagnosis. Performance of these algorithms has been hindered by the scarcity of annotated data, especially those with fine-grained delineation of the thromboembolic burden. In this paper we attempt to address this issue by introducing a weakly supervised learning pipeline, that leverages model explainability to generate fine-grained (pixel level) masks for embolisms starting from more coarse-grained (binary, image level) PE annotations. Furthermore, we show that training models using the automatically generated pixel annotations yields good PE localization performance. We demonstrate the effectiveness of our pipeline on the large-scale, multi-center RSPECT augmented dataset for PE detection and localization.
☆ Temporal Linear Item-Item Model for Sequential Recommendation WSDM 2025
In sequential recommendation (SR), neural models have been actively explored due to their remarkable performance, but they suffer from inefficiency inherent to their complexity. On the other hand, linear SR models exhibit high efficiency and achieve competitive or superior accuracy compared to neural models. However, they solely deal with the sequential order of items (i.e., sequential information) and overlook the actual timestamp (i.e., temporal information). It is limited to effectively capturing various user preference drifts over time. To address this issue, we propose a novel linear SR model, named TemporAl LinEar item-item model (TALE), incorporating temporal information while preserving training/inference efficiency, with three key components. (i) Single-target augmentation concentrates on a single target item, enabling us to learn the temporal correlation for the target item. (ii) Time interval-aware weighting utilizes the actual timestamp to discern the item correlation depending on time intervals. (iii) Trend-aware normalization reflects the dynamic shift of item popularity over time. Our empirical studies show that TALE outperforms ten competing SR models by up to 18.71% gains on five benchmark datasets. It also exhibits remarkable effectiveness in evaluating long-tail items by up to 30.45% gains. The source code is available at https://github.com/psm1206/TALE.
comment: Accepted by WSDM 2025
☆ A Spectral Framework for Tracking Communities in Evolving Networks
Discovering and tracking communities in time-varying networks is an important task in network science, motivated by applications in fields ranging from neuroscience to sociology. In this work, we characterize the celebrated family of spectral methods for static clustering in terms of the low-rank approximation of high-dimensional node embeddings. From this perspective, it becomes natural to view the evolving community detection problem as one of subspace tracking on the Grassmann manifold. While the resulting optimization problem is nonconvex, we adopt a block majorize-minimize Riemannian optimization scheme to learn the Grassmann geodesic which best fits the data. Our framework generalizes any static spectral community detection approach and leads to algorithms achieving favorable performance on synthetic and real temporal networks, including those that are weighted, signed, directed, mixed-membership, multiview, hierarchical, cocommunity-structured, bipartite, or some combination thereof. We demonstrate how to specifically cast a wide variety of methods into our framework, and demonstrate greatly improved dynamic community detection results in all cases.
comment: 34 pages, 13 figures
☆ NeSyA: Neurosymbolic Automata
Neurosymbolic Artificial Intelligence (NeSy) has emerged as a promising direction to integrate low level perception with high level reasoning. Unfortunately, little attention has been given to developing NeSy systems tailored to temporal/sequential problems. This entails reasoning symbolically over sequences of subsymbolic observations towards a target prediction. We show that using a probabilistic semantics symbolic automata, which combine the power of automata for temporal structure specification with that of propositional logic, can be used to reason efficiently and differentiably over subsymbolic sequences. The proposed system, which we call NeSyA (Neuro Symbolic Automata), is shown to either scale or perform better than existing NeSy approaches when applied to problems with a temporal component.
☆ Addressing Key Challenges of Adversarial Attacks and Defenses in the Tabular Domain: A Methodological Framework for Coherence and Consistency
Machine learning models trained on tabular data are vulnerable to adversarial attacks, even in realistic scenarios where attackers have access only to the model's outputs. Researchers evaluate such attacks by considering metrics like success rate, perturbation magnitude, and query count. However, unlike other data domains, the tabular domain contains complex interdependencies among features, presenting a unique aspect that should be evaluated: the need for the attack to generate coherent samples and ensure feature consistency for indistinguishability. Currently, there is no established methodology for evaluating adversarial samples based on these criteria. In this paper, we address this gap by proposing new evaluation criteria tailored for tabular attacks' quality; we defined anomaly-based framework to assess the distinguishability of adversarial samples and utilize the SHAP explainability technique to identify inconsistencies in the model's decision-making process caused by adversarial samples. These criteria could form the basis for potential detection methods and be integrated into established evaluation metrics for assessing attack's quality Additionally, we introduce a novel technique for perturbing dependent features while maintaining coherence and feature consistency within the sample. We compare different attacks' strategies, examining black-box query-based attacks and transferability-based gradient attacks across four target models. Our experiments, conducted on benchmark tabular datasets, reveal significant differences between the examined attacks' strategies in terms of the attacker's risk and effort and the attacks' quality. The findings provide valuable insights on the strengths, limitations, and trade-offs of various adversarial attacks in the tabular domain, laying a foundation for future research on attacks and defense development.
☆ Label Distribution Learning using the Squared Neural Family on the Probability Simplex
Label distribution learning (LDL) provides a framework wherein a distribution over categories rather than a single category is predicted, with the aim of addressing ambiguity in labeled data. Existing research on LDL mainly focuses on the task of point estimation, i.e., pinpointing an optimal distribution in the probability simplex conditioned on the input sample. In this paper, we estimate a probability distribution of all possible label distributions over the simplex, by unleashing the expressive power of the recently introduced Squared Neural Family (SNEFY). With the modeled distribution, label distribution prediction can be achieved by performing the expectation operation to estimate the mean of the distribution of label distributions. Moreover, more information about the label distribution can be inferred, such as the prediction reliability and uncertainties. We conduct extensive experiments on the label distribution prediction task, showing that our distribution modeling based method can achieve very competitive label distribution prediction performance compared with the state-of-the-art baselines. Additional experiments on active learning and ensemble learning demonstrate that our probabilistic approach can effectively boost the performance in these settings, by accurately estimating the prediction reliability and uncertainties.
☆ ConceptSearch: Towards Efficient Program Search Using LLMs for Abstraction and Reasoning Corpus (ARC) AAAI 2025
The Abstraction and Reasoning Corpus (ARC) poses a significant challenge to artificial intelligence, demanding broad generalization and few-shot learning capabilities that remain elusive for current deep learning methods, including large language models (LLMs). While LLMs excel in program synthesis, their direct application to ARC yields limited success. To address this, we introduce ConceptSearch, a novel function-search algorithm that leverages LLMs for program generation and employs a concept-based scoring method to guide the search efficiently. Unlike simplistic pixel-based metrics like Hamming distance, ConceptSearch evaluates programs on their ability to capture the underlying transformation concept reflected in the input-output examples. We explore three scoring functions: Hamming distance, a CNN-based scoring function, and an LLM-based natural language scoring function. Experimental results demonstrate the effectiveness of ConceptSearch, achieving a significant performance improvement over direct prompting with GPT-4. Moreover, our novel concept-based scoring exhibits up to 30% greater efficiency compared to Hamming distance, measured in terms of the number of iterations required to reach the correct solution. These findings highlight the potential of LLM-driven program search when integrated with concept-based guidance for tackling challenging generalization problems like ARC. Code: https://github.com/kksinghal/concept-search
comment: 8 pages, 7 figures, to appear at AAAI 2025
☆ High-dimensional classification problems with Barron regular boundaries under margin conditions
We prove that a classifier with a Barron-regular decision boundary can be approximated with a rate of high polynomial degree by ReLU neural networks with three hidden layers when a margin condition is assumed. In particular, for strong margin conditions, high-dimensional discontinuous classifiers can be approximated with a rate that is typically only achievable when approximating a low-dimensional smooth function. We demonstrate how these expression rate bounds imply fast-rate learning bounds that are close to $n^{-1}$ where $n$ is the number of samples. In addition, we carry out comprehensive numerical experimentation on binary classification problems with various margins. We study three different dimensions, with the highest dimensional problem corresponding to images from the MNIST data set.
☆ HARP: Hesitation-Aware Reframing in Transformer Inference Pass
This paper aims to improve the performance of large language models by addressing the variable computational demands in inference steps, where some tokens require more computational resources than others. We present HARP, a simple modification to "off-the-shelf" Transformer forward pass. Drawing from hesitation and the framing effect in decision-making, HARP selectively applies additional computation when the model encounters uncertainty during token generation. Our method mimics human cognitive processes by pausing at difficult decision points and reframing inputs for a different perspective. Unlike other approaches, HARP is model-agnostic, training-free, and easy to implement. We thoroughly evaluate our method across various downstream tasks and model sizes, demonstrating performance improvements up to +5.16%. Notably, HARP achieves these gains while maintaining inference times twice faster than beam search. Simple and yet with significant gains, HARP offers a practical solution for enhancing the performance of Transformer-based language models with minimal computational impact.
☆ Temporal-Aware Evaluation and Learning for Temporal Graph Neural Networks
Temporal Graph Neural Networks (TGNNs) are a family of graph neural networks designed to model and learn dynamic information from temporal graphs. Given their substantial empirical success, there is an escalating interest in TGNNs within the research community. However, the majority of these efforts have been channelled towards algorithm and system design, with the evaluation metrics receiving comparatively less attention. Effective evaluation metrics are crucial for providing detailed performance insights, particularly in the temporal domain. This paper investigates the commonly used evaluation metrics for TGNNs and illustrates the failure mechanisms of these metrics in capturing essential temporal structures in the predictive behaviour of TGNNs. We provide a mathematical formulation of existing performance metrics and utilize an instance-based study to underscore their inadequacies in identifying volatility clustering (the occurrence of emerging errors within a brief interval). This phenomenon has profound implications for both algorithm and system design in the temporal domain. To address this deficiency, we introduce a new volatility-aware evaluation metric (termed volatility cluster statistics), designed for a more refined analysis of model temporal performance. Additionally, we demonstrate how this metric can serve as a temporal-volatility-aware training objective to alleviate the clustering of temporal errors. Through comprehensive experiments on various TGNN models, we validate our analysis and the proposed approach. The empirical results offer revealing insights: 1) existing TGNNs are prone to making errors with volatility clustering, and 2) TGNNs with different mechanisms to capture temporal information exhibit distinct volatility clustering patterns. Our empirical findings demonstrate that our proposed training objective effectively reduces volatility clusters in error.
☆ PTSBench: A Comprehensive Post-Training Sparsity Benchmark Towards Algorithms and Models
With the increased attention to model efficiency, post-training sparsity (PTS) has become more and more prevalent because of its effectiveness and efficiency. However, there remain questions on better practice of PTS algorithms and the sparsification ability of models, which hinders the further development of this area. Therefore, a benchmark to comprehensively investigate the issues above is urgently needed. In this paper, we propose the first comprehensive post-training sparsity benchmark called PTSBench towards algorithms and models. We benchmark 10+ PTS general-pluggable fine-grained techniques on 3 typical tasks using over 40 off-the-shelf model architectures. Through extensive experiments and analyses, we obtain valuable conclusions and provide several insights from both algorithms and model aspects. Our PTSBench can provide (1) new observations for a better understanding of the PTS algorithms, (2) in-depth and comprehensive evaluations for the sparsification ability of models, and (3) a well-structured and easy-integrate open-source framework. We hope this work will provide illuminating conclusions and advice for future studies of post-training sparsity methods and sparsification-friendly model design. The code for our PTSBench is released at \href{https://github.com/ModelTC/msbench}{https://github.com/ModelTC/msbench}.
☆ Modeling High-Resolution Spatio-Temporal Wind with Deep Echo State Networks and Stochastic Partial Differential Equations
In the past decades, clean and renewable energy has gained increasing attention due to a global effort on carbon footprint reduction. In particular, Saudi Arabia is gradually shifting its energy portfolio from an exclusive use of oil to a reliance on renewable energy, and, in particular, wind. Modeling wind for assessing potential energy output in a country as large, geographically diverse and understudied as Saudi Arabia is a challenge which implies highly non-linear dynamic structures in both space and time. To address this, we propose a spatio-temporal model whose spatial information is first reduced via an energy distance-based approach and then its dynamical behavior is informed by a sparse and stochastic recurrent neural network (Echo State Network). Finally, the full spatial data is reconstructed by means of a non-stationary stochastic partial differential equation-based approach. Our model can capture the fine scale wind structure and produce more accurate forecasts of both wind speed and energy in lead times of interest for energy grid management and save annually as much as one million dollar against the closest competitive model.
☆ MemHunter: Automated and Verifiable Memorization Detection at Dataset-scale in LLMs
Large language models (LLMs) have been shown to memorize and reproduce content from their training data, raising significant privacy concerns, especially with web-scale datasets. Existing methods for detecting memorization are largely sample-specific, relying on manually crafted or discretely optimized memory-inducing prompts generated on a per-sample basis, which become impractical for dataset-level detection due to the prohibitive computational cost of iterating over all samples. In real-world scenarios, data owners may need to verify whether a susceptible LLM has memorized their dataset, particularly if the LLM may have collected the data from the web without authorization. To address this, we introduce \textit{MemHunter}, which trains a memory-inducing LLM and employs hypothesis testing to efficiently detect memorization at the dataset level, without requiring sample-specific memory inducing. Experiments on models such as Pythia and Llama-2 demonstrate that \textit{MemHunter} can extract up to 40\% more training data than existing methods under constrained time resources and reduce search time by up to 80\% when integrated as a plug-in. Crucially, \textit{MemHunter} is the first method capable of dataset-level memorization detection, providing an indispensable tool for assessing privacy risks in LLMs that are powered by vast web-sourced datasets.
☆ Buster: Incorporating Backdoor Attacks into Text Encoder to Mitigate NSFW Content Generation
In the digital age, the proliferation of deep learning models has led to significant concerns about the generation of Not Safe for Work (NSFW) content. Existing defense methods primarily involve model fine-tuning and post-hoc content moderation. However, these approaches often lack scalability in eliminating harmful content, degrade the quality of benign image generation, or incur high inference costs. To tackle these challenges, we propose an innovative framework called \textbf{Buster}, which injects backdoor attacks into the text encoder to prevent NSFW content generation. Specifically, Buster leverages deep semantic information rather than explicit prompts as triggers, redirecting NSFW prompts towards targeted benign prompts. This approach demonstrates exceptional resilience and scalability in mitigating NSFW content. Remarkably, Buster fine-tunes the text encoder of Text-to-Image models within just five minutes, showcasing high efficiency. Our extensive experiments reveal that Buster outperforms all other baselines, achieving superior NSFW content removal rate while preserving the quality of harmless images.
☆ Developing a Dataset-Adaptive, Normalized Metric for Machine Learning Model Assessment: Integrating Size, Complexity, and Class Imbalance
Traditional metrics like accuracy, F1-score, and precision are frequently used to evaluate machine learning models, however they may not be sufficient for evaluating performance on tiny, unbalanced, or high-dimensional datasets. A dataset-adaptive, normalized metric that incorporates dataset characteristics like size, feature dimensionality, class imbalance, and signal-to-noise ratio is presented in this study. Early insights into the model's performance potential in challenging circumstances are provided by the suggested metric, which offers a scalable and adaptable evaluation framework. The metric's capacity to accurately forecast model scalability and performance is demonstrated via experimental validation spanning classification, regression, and clustering tasks, guaranteeing solid assessments in settings with limited data. This method has important ramifications for effective resource allocation and model optimization in machine learning workflows.
comment: 36 pages, 17 figures. Includes results validated on datasets from UCI Machine Learning Repository
☆ A Dynamical Systems-Inspired Pruning Strategy for Addressing Oversmoothing in Graph Neural Networks
Oversmoothing in Graph Neural Networks (GNNs) poses a significant challenge as network depth increases, leading to homogenized node representations and a loss of expressiveness. In this work, we approach the oversmoothing problem from a dynamical systems perspective, providing a deeper understanding of the stability and convergence behavior of GNNs. Leveraging insights from dynamical systems theory, we identify the root causes of oversmoothing and propose \textbf{\textit{DYNAMO-GAT}}. This approach utilizes noise-driven covariance analysis and Anti-Hebbian principles to selectively prune redundant attention weights, dynamically adjusting the network's behavior to maintain node feature diversity and stability. Our theoretical analysis reveals how DYNAMO-GAT disrupts the convergence to oversmoothed states, while experimental results on benchmark datasets demonstrate its superior performance and efficiency compared to traditional and state-of-the-art methods. DYNAMO-GAT not only advances the theoretical understanding of oversmoothing through the lens of dynamical systems but also provides a practical and effective solution for improving the stability and expressiveness of deep GNNs.
comment: 26 pages
☆ Optimization Can Learn Johnson Lindenstrauss Embeddings
Embeddings play a pivotal role across various disciplines, offering compact representations of complex data structures. Randomized methods like Johnson-Lindenstrauss (JL) provide state-of-the-art and essentially unimprovable theoretical guarantees for achieving such representations. These guarantees are worst-case and in particular, neither the analysis, nor the algorithm, takes into account any potential structural information of the data. The natural question is: must we randomize? Could we instead use an optimization-based approach, working directly with the data? A first answer is no: as we show, the distance-preserving objective of JL has a non-convex landscape over the space of projection matrices, with many bad stationary points. But this is not the final answer. We present a novel method motivated by diffusion models, that circumvents this fundamental challenge: rather than performing optimization directly over the space of projection matrices, we use optimization over the larger space of random solution samplers, gradually reducing the variance of the sampler. We show that by moving through this larger space, our objective converges to a deterministic (zero variance) solution, avoiding bad stationary points. This method can also be seen as an optimization-based derandomization approach and is an idea and method that we believe can be applied to many other problems.
☆ Human-Computer Interaction and Human-AI Collaboration in Advanced Air Mobility: A Comprehensive Review
The increasing rates of global urbanization and vehicle usage are leading to a shift of mobility to the third dimension-through Advanced Air Mobility (AAM)-offering a promising solution for faster, safer, cleaner, and more efficient transportation. As air transportation continues to evolve with more automated and autonomous systems, advancements in AAM require a deep understanding of human-computer interaction and human-AI collaboration to ensure safe and effective operations in complex urban and regional environments. There has been a significant increase in publications regarding these emerging applications; thus, there is a need to review developments in this area. This paper comprehensively reviews the current state of research on human-computer interaction and human-AI collaboration in AAM. Specifically, we focus on AAM applications related to the design of human-machine interfaces for various uses, including pilot training, air traffic management, and the integration of AI-assisted decision-making systems with immersive technologies such as extended, virtual, mixed, and augmented reality devices. Additionally, we provide a comprehensive analysis of the challenges AAM encounters in integrating human-computer frameworks, including unique challenges associated with these interactions, such as trust in AI systems and safety concerns. Finally, we highlight emerging opportunities and propose future research directions to bridge the gap between human factors and technological advancements in AAM.
☆ CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding
Electroencephalography (EEG) is a non-invasive technique to measure and record brain electrical activity, widely used in various BCI and healthcare applications. Early EEG decoding methods rely on supervised learning, limited by specific tasks and datasets, hindering model performance and generalizability. With the success of large language models, there is a growing body of studies focusing on EEG foundation models. However, these studies still leave challenges: Firstly, most of existing EEG foundation models employ full EEG modeling strategy. It models the spatial and temporal dependencies between all EEG patches together, but ignores that the spatial and temporal dependencies are heterogeneous due to the unique structural characteristics of EEG signals. Secondly, existing EEG foundation models have limited generalizability on a wide range of downstream BCI tasks due to varying formats of EEG data, making it challenging to adapt to. To address these challenges, we propose a novel foundation model called CBraMod. Specifically, we devise a criss-cross transformer as the backbone to thoroughly leverage the structural characteristics of EEG signals, which can model spatial and temporal dependencies separately through two parallel attention mechanisms. And we utilize an asymmetric conditional positional encoding scheme which can encode positional information of EEG patches and be easily adapted to the EEG with diverse formats. CBraMod is pre-trained on a very large corpus of EEG through patch-based masked EEG reconstruction. We evaluate CBraMod on up to 10 downstream BCI tasks (12 public datasets). CBraMod achieves the state-of-the-art performance across the wide range of tasks, proving its strong capability and generalizability. The source code is publicly available at \url{https://github.com/wjq-learning/CBraMod}.
☆ Adversarial Filtering Based Evasion and Backdoor Attacks to EEG-Based Brain-Computer Interfaces
A brain-computer interface (BCI) enables direct communication between the brain and an external device. Electroencephalogram (EEG) is a common input signal for BCIs, due to its convenience and low cost. Most research on EEG-based BCIs focuses on the accurate decoding of EEG signals, while ignoring their security. Recent studies have shown that machine learning models in BCIs are vulnerable to adversarial attacks. This paper proposes adversarial filtering based evasion and backdoor attacks to EEG-based BCIs, which are very easy to implement. Experiments on three datasets from different BCI paradigms demonstrated the effectiveness of our proposed attack approaches. To our knowledge, this is the first study on adversarial filtering for EEG-based BCIs, raising a new security concern and calling for more attention on the security of BCIs.
☆ Moderating the Generalization of Score-based Generative Model
Score-based Generative Models (SGMs) have demonstrated remarkable generalization abilities, e.g. generating unseen, but natural data. However, the greater the generalization power, the more likely the unintended generalization, and the more dangerous the abuse. Research on moderated generalization in SGMs remains limited. To fill this gap, we first examine the current 'gold standard' in Machine Unlearning (MU), i.e., re-training the model after removing the undesirable training data, and find it does not work in SGMs. Further analysis of score functions reveals that the MU 'gold standard' does not alter the original score function, which explains its ineffectiveness. Based on this insight, we propose the first Moderated Score-based Generative Model (MSGM), which introduces a novel score adjustment strategy that redirects the score function away from undesirable data during the continuous-time stochastic differential equation process. Extensive experimental results demonstrate that MSGM significantly reduces the likelihood of generating undesirable content while preserving high visual quality for normal image generation. Albeit designed for SGMs, MSGM is a general and flexible MU framework that is compatible with diverse diffusion architectures (SGM and DDPM) and training strategies (re-training and fine-tuning), and enables zero-shot transfer of the pre-trained models to downstream tasks, e.g. image inpainting and reconstruction. The code will be shared upon acceptance.
☆ T-TIME: Test-Time Information Maximization Ensemble for Plug-and-Play BCIs
Objective: An electroencephalogram (EEG)-based brain-computer interface (BCI) enables direct communication between the human brain and a computer. Due to individual differences and non-stationarity of EEG signals, such BCIs usually require a subject-specific calibration session before each use, which is time-consuming and user-unfriendly. Transfer learning (TL) has been proposed to shorten or eliminate this calibration, but existing TL approaches mainly consider offline settings, where all unlabeled EEG trials from the new user are available. Methods: This paper proposes Test-Time Information Maximization Ensemble (T-TIME) to accommodate the most challenging online TL scenario, where unlabeled EEG data from the new user arrive in a stream, and immediate classification is performed. T-TIME initializes multiple classifiers from the aligned source data. When an unlabeled test EEG trial arrives, T-TIME first predicts its labels using ensemble learning, and then updates each classifier by conditional entropy minimization and adaptive marginal distribution regularization. Our code is publicized. Results: Extensive experiments on three public motor imagery based BCI datasets demonstrated that T-TIME outperformed about 20 classical and state-of-the-art TL approaches. Significance: To our knowledge, this is the first work on test time adaptation for calibration-free EEG-based BCIs, making plug-and-play BCIs possible.
☆ Parseval Regularization for Continual Reinforcement Learning
Loss of plasticity, trainability loss, and primacy bias have been identified as issues arising when training deep neural networks on sequences of tasks -- all referring to the increased difficulty in training on new tasks. We propose to use Parseval regularization, which maintains orthogonality of weight matrices, to preserve useful optimization properties and improve training in a continual reinforcement learning setting. We show that it provides significant benefits to RL agents on a suite of gridworld, CARL and MetaWorld tasks. We conduct comprehensive ablations to identify the source of its benefits and investigate the effect of certain metrics associated to network trainability including weight matrix rank, weight norms and policy entropy.
☆ A Consolidated Volatility Prediction with Back Propagation Neural Network and Genetic Algorithm ICML 2024
This paper provides a unique approach with AI algorithms to predict emerging stock markets volatility. Traditionally, stock volatility is derived from historical volatility,Monte Carlo simulation and implied volatility as well. In this paper, the writer designs a consolidated model with back-propagation neural network and genetic algorithm to predict future volatility of emerging stock markets and found that the results are quite accurate with low errors.
comment: 6 pages, 7 figures, 1 table, The paper will be published by IEEE on conference: 2024 3rd International Conference on Image Processing, Computer Vision and Machine Learning (ICICML 2024)
☆ Taylor Outlier Exposure
Out-of-distribution (OOD) detection is the task of identifying data sampled from distributions that were not used during training. This task is essential for reliable machine learning and a better understanding of their generalization capabilities. Among OOD detection methods, Outlier Exposure (OE) significantly enhances OOD detection performance and generalization ability by exposing auxiliary OOD data to the model. However, constructing clean auxiliary OOD datasets, uncontaminated by in-distribution (ID) samples, is essential for OE; generally, a noisy OOD dataset contaminated with ID samples negatively impacts OE training dynamics and final detection performance. Furthermore, as dataset scale increases, constructing clean OOD data becomes increasingly challenging and costly. To address these challenges, we propose Taylor Outlier Exposure (TaylorOE), an OE-based approach with regularization that allows training on noisy OOD datasets contaminated with ID samples. Specifically, we represent the OE regularization term as a polynomial function via a Taylor expansion, allowing us to control the regularization strength for ID data in the auxiliary OOD dataset by adjusting the order of Taylor expansion. In our experiments on the OOD detection task with clean and noisy OOD datasets, we demonstrate that the proposed method consistently outperforms conventional methods and analyze our regularization term to show its effectiveness. Our implementation code of TaylorOE is available at \url{https://github.com/fukuchan41/TaylorOE}.
☆ Incremental Gaussian Mixture Clustering for Data Streams
The problem of analyzing data streams of very large volumes is important and is very desirable for many application domains. In this paper we present and demonstrate effective working of an algorithm to find clusters and anomalous data points in a streaming datasets. Entropy minimization is used as a criterion for defining and updating clusters formed from a streaming dataset. As the clusters are formed we also identify anomalous datapoints that show up far away from all known clusters. With a number of 2-D datasets we demonstrate the effectiveness of discovering the clusters and also identifying anomalous data points.
☆ Learnable Sparse Customization in Heterogeneous Edge Computing ICDE 2025
To effectively manage and utilize massive distributed data at the network edge, Federated Learning (FL) has emerged as a promising edge computing paradigm across data silos. However, FL still faces two challenges: system heterogeneity (i.e., the diversity of hardware resources across edge devices) and statistical heterogeneity (i.e., non-IID data). Although sparsification can extract diverse submodels for diverse clients, most sparse FL works either simply assign submodels with artificially-given rigid rules or prune partial parameters using heuristic strategies, resulting in inflexible sparsification and poor performance. In this work, we propose Learnable Personalized Sparsification for heterogeneous Federated learning (FedLPS), which achieves the learnable customization of heterogeneous sparse models with importance-associated patterns and adaptive ratios to simultaneously tackle system and statistical heterogeneity. Specifically, FedLPS learns the importance of model units on local data representation and further derives an importance-based sparse pattern with minimal heuristics to accurately extract personalized data features in non-IID settings. Furthermore, Prompt Upper Confidence Bound Variance (P-UCBV) is designed to adaptively determine sparse ratios by learning the superimposed effect of diverse device capabilities and non-IID data, aiming at resource self-adaptation with promising accuracy. Extensive experiments show that FedLPS outperforms status quo approaches in accuracy and training costs, which improves accuracy by 1.28%-59.34% while reducing running time by more than 68.80%.
comment: Accepted by ICDE 2025
☆ MAPLE: A Framework for Active Preference Learning Guided by Large Language Models
The advent of large language models (LLMs) has sparked significant interest in using natural language for preference learning. However, existing methods often suffer from high computational burdens, taxing human supervision, and lack of interpretability. To address these issues, we introduce MAPLE, a framework for large language model-guided Bayesian active preference learning. MAPLE leverages LLMs to model the distribution over preference functions, conditioning it on both natural language feedback and conventional preference learning feedback, such as pairwise trajectory rankings. MAPLE also employs active learning to systematically reduce uncertainty in this distribution and incorporates a language-conditioned active query selection mechanism to identify informative and easy-to-answer queries, thus reducing human burden. We evaluate MAPLE's sample efficiency and preference inference quality across two benchmarks, including a real-world vehicle route planning benchmark using OpenStreetMap data. Our results demonstrate that MAPLE accelerates the learning process and effectively improves humans' ability to answer queries.
☆ Crack-EdgeSAM Self-Prompting Crack Segmentation System for Edge Devices
Structural health monitoring (SHM) is essential for the early detection of infrastructure defects, such as cracks in concrete bridge pier. but often faces challenges in efficiency and accuracy in complex environments. Although the Segment Anything Model (SAM) achieves excellent segmentation performance, its computational demands limit its suitability for real-time applications on edge devices. To address these challenges, this paper proposes Crack-EdgeSAM, a self-prompting crack segmentation system that integrates YOLOv8 for generating prompt boxes and a fine-tuned EdgeSAM model for crack segmentation. To ensure computational efficiency, the method employs ConvLoRA, a Parameter-Efficient Fine-Tuning (PEFT) technique, along with DiceFocalLoss to fine-tune the EdgeSAM model. Our experimental results on public datasets and the climbing robot automatic inspections demonstrate that the system achieves high segmentation accuracy and significantly enhanced inference speed compared to the most recent methods. Notably, the system processes 1024 x 1024 pixels images at 46 FPS on our PC and 8 FPS on Jetson Orin Nano.
☆ Hierarchical Split Federated Learning: Convergence Analysis and System Optimization
As AI models expand in size, it has become increasingly challenging to deploy federated learning (FL) on resource-constrained edge devices. To tackle this issue, split federated learning (SFL) has emerged as an FL framework with reduced workload on edge devices via model splitting; it has received extensive attention from the research community in recent years. Nevertheless, most prior works on SFL focus only on a two-tier architecture without harnessing multi-tier cloudedge computing resources. In this paper, we intend to analyze and optimize the learning performance of SFL under multi-tier systems. Specifically, we propose the hierarchical SFL (HSFL) framework and derive its convergence bound. Based on the theoretical results, we formulate a joint optimization problem for model splitting (MS) and model aggregation (MA). To solve this rather hard problem, we then decompose it into MS and MA subproblems that can be solved via an iterative descending algorithm. Simulation results demonstrate that the tailored algorithm can effectively optimize MS and MA for SFL within virtually any multi-tier system.
comment: 15 pages, 9 figures
☆ A Progressive Image Restoration Network for High-order Degradation Imaging in Remote Sensing
Recently, deep learning methods have gained remarkable achievements in the field of image restoration for remote sensing (RS). However, most existing RS image restoration methods focus mainly on conventional first-order degradation models, which may not effectively capture the imaging mechanisms of remote sensing images. Furthermore, many RS image restoration approaches that use deep learning are often criticized for their lacks of architecture transparency and model interpretability. To address these problems, we propose a novel progressive restoration network for high-order degradation imaging (HDI-PRNet), to progressively restore different image degradation. HDI-PRNet is developed based on the theoretical framework of degradation imaging, offering the benefit of mathematical interpretability within the unfolding network. The framework is composed of three main components: a module for image denoising that relies on proximal mapping prior learning, a module for image deblurring that integrates Neumann series expansion with dual-domain degradation learning, and a module for super-resolution. Extensive experiments demonstrate that our method achieves superior performance on both synthetic and real remote sensing images.
comment: 14 pages
☆ Epidemiological Model Calibration via Graybox Bayesian Optimization
In this study, we focus on developing efficient calibration methods via Bayesian decision-making for the family of compartmental epidemiological models. The existing calibration methods usually assume that the compartmental model is cheap in terms of its output and gradient evaluation, which may not hold in practice when extending them to more general settings. Therefore, we introduce model calibration methods based on a "graybox" Bayesian optimization (BO) scheme, more efficient calibration for general epidemiological models. This approach uses Gaussian processes as a surrogate to the expensive model, and leverages the functional structure of the compartmental model to enhance calibration performance. Additionally, we develop model calibration methods via a decoupled decision-making strategy for BO, which further exploits the decomposable nature of the functional structure. The calibration efficiencies of the multiple proposed schemes are evaluated based on various data generated by a compartmental model mimicking real-world epidemic processes, and real-world COVID-19 datasets. Experimental results demonstrate that our proposed graybox variants of BO schemes can efficiently calibrate computationally expensive models and further improve the calibration performance measured by the logarithm of mean square errors and achieve faster performance convergence in terms of BO iterations. We anticipate that the proposed calibration methods can be extended to enable fast calibration of more complex epidemiological models, such as the agent-based models.
☆ PrisonBreak: Jailbreaking Large Language Models with Fewer Than Twenty-Five Targeted Bit-flips
We introduce a new class of attacks on commercial-scale (human-aligned) language models that induce jailbreaking through targeted bitwise corruptions in model parameters. Our adversary can jailbreak billion-parameter language models with fewer than 25 bit-flips in all cases$-$and as few as 5 in some$-$using up to 40$\times$ less bit-flips than existing attacks on computer vision models at least 100$\times$ smaller. Unlike prompt-based jailbreaks, our attack renders these models in memory 'uncensored' at runtime, allowing them to generate harmful responses without any input modifications. Our attack algorithm efficiently identifies target bits to flip, offering up to 20$\times$ more computational efficiency than previous methods. This makes it practical for language models with billions of parameters. We show an end-to-end exploitation of our attack using software-induced fault injection, Rowhammer (RH). Our work examines 56 DRAM RH profiles from DDR4 and LPDDR4X devices with different RH vulnerabilities. We show that our attack can reliably induce jailbreaking in systems similar to those affected by prior bit-flip attacks. Moreover, our approach remains effective even against highly RH-secure systems (e.g., 46$\times$ more secure than previously tested systems). Our analyses further reveal that: (1) models with less post-training alignment require fewer bit flips to jailbreak; (2) certain model components, such as value projection layers, are substantially more vulnerable than others; and (3) our method is mechanistically different than existing jailbreaks. Our findings highlight a pressing, practical threat to the language model ecosystem and underscore the need for research to protect these models from bit-flip attacks.
♻ ☆ Impact of Leakage on Data Harmonization in Machine Learning Pipelines in Class Imbalance Across Sites
Machine learning (ML) models benefit from large datasets. Collecting data in biomedical domains is costly and challenging, hence, combining datasets has become a common practice. However, datasets obtained under different conditions could present undesired site-specific variability. Data harmonization methods aim to remove site-specific variance while retaining biologically relevant information. This study evaluates the effectiveness of popularly used ComBat-based methods for harmonizing data in scenarios where the class balance is not equal across sites. We find that these methods struggle with data leakage issues. To overcome this problem, we propose a novel approach PrettYharmonize, designed to harmonize data by pretending the target labels. We validate our approach using controlled datasets designed to benchmark the utility of harmonization. Finally, using real-world MRI and clinical data, we compare leakage-prone methods with PrettYharmonize and show that it achieves comparable performance while avoiding data leakage, particularly in site-target-dependence scenarios.
♻ ☆ Right on Time: Revising Time Series Models by Constraining their Explanations
The reliability of deep time series models is often compromised by their tendency to rely on confounding factors, which may lead to incorrect outputs. Our newly recorded, naturally confounded dataset named P2S from a real mechanical production line emphasizes this. To avoid "Clever-Hans" moments in time series, i.e., to mitigate confounders, we introduce the method Right on Time (RioT). RioT enables, for the first time, interactions with model explanations across both the time and frequency domain. Feedback on explanations in both domains is then used to constrain the model, steering it away from the annotated confounding factors. The dual-domain interaction strategy is crucial for effectively addressing confounders in time series datasets. We empirically demonstrate that RioT can effectively guide models away from the wrong reasons in P2S as well as popular time series classification and forecasting datasets.
♻ ☆ Beyond Retrieval: Generating Narratives in Conversational Recommender Systems
The recent advances in Large Language Model's generation and reasoning capabilities present an opportunity to develop truly conversational recommendation systems. However, effectively integrating recommender system knowledge into LLMs for natural language generation which is tailored towards recommendation tasks remains a challenge. This paper addresses this challenge by making two key contributions. First, we introduce a new dataset (REGEN) for natural language generation tasks in conversational recommendations. REGEN (Reviews Enhanced with GEnerative Narratives) extends the Amazon Product Reviews dataset with rich user narratives, including personalized explanations of product preferences, product endorsements for recommended items, and summaries of user purchase history. REGEN is made publicly available to facilitate further research. Furthermore, we establish benchmarks using well-known generative metrics, and perform an automated evaluation of the new dataset using a rater LLM. Second, the paper introduces a fusion architecture (CF model with an LLM) which serves as a baseline for REGEN. And to the best of our knowledge, represents the first attempt to analyze the capabilities of LLMs in understanding recommender signals and generating rich narratives. We demonstrate that LLMs can effectively learn from simple fusion architectures utilizing interaction-based CF embeddings, and this can be further enhanced using the metadata and personalization data associated with items. Our experiments show that combining CF and content embeddings leads to improvements of 4-12% in key language metrics compared to using either type of embedding individually. We also provide an analysis to interpret how CF and content embeddings contribute to this new generative task.
♻ ☆ A tutorial on automatic differentiation with complex numbers
Automatic differentiation is everywhere, but there exists only minimal documentation of how it works in complex arithmetic beyond stating "derivatives in $\mathbb{C}^d$" $\cong$ "derivatives in $\mathbb{R}^{2d}$" and, at best, shallow references to Wirtinger calculus. Unfortunately, the equivalence $\mathbb{C}^d \cong \mathbb{R}^{2d}$ becomes insufficient as soon as we need to derive custom gradient rules, e.g., to avoid differentiating "through" expensive linear algebra functions or differential equation simulators. To combat such a lack of documentation, this article surveys forward- and reverse-mode automatic differentiation with complex numbers, covering topics such as Wirtinger derivatives, a modified chain rule, and different gradient conventions while explicitly avoiding holomorphicity and the Cauchy--Riemann equations (which would be far too restrictive). To be precise, we will derive, explain, and implement a complex version of Jacobian-vector and vector-Jacobian products almost entirely with linear algebra without relying on complex analysis or differential geometry. This tutorial is a call to action, for users and developers alike, to take complex values seriously when implementing custom gradient propagation rules -- the manuscript explains how.
♻ ☆ An objective function for order preserving hierarchical clustering
We present a theory and an objective function for similarity-based hierarchical clustering of probabilistic partial orders and directed acyclic graphs (DAGs). Specifically, given elements $x \le y$ in the partial order, and their respective clusters $[x]$ and $[y]$, the theory yields an order relation $\le'$ on the clusters such that $[x]\le'[y]$. The theory provides a concise definition of order-preserving hierarchical clustering, and offers a classification theorem identifying the order-preserving trees (dendrograms). To determine the optimal order-preserving trees, we develop an objective function that frames the problem as a bi-objective optimisation, aiming to satisfy both the order relation and the similarity measure. We prove that the optimal trees under the objective are both order-preserving and exhibit high-quality hierarchical clustering. Since finding an optimal solution is NP-hard, we introduce a polynomial-time approximation algorithm and demonstrate that the method outperforms existing methods for order-preserving hierarchical clustering by a significant margin.
comment: 39 pages
♻ ☆ The BrowserGym Ecosystem for Web Agent Research
The BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents, particularly those leveraging automation and Large Language Models (LLMs) for web interaction tasks. Many existing benchmarks suffer from fragmentation and inconsistent evaluation methodologies, making it challenging to achieve reliable comparisons and reproducible results. BrowserGym aims to solve this by providing a unified, gym-like environment with well-defined observation and action spaces, facilitating standardized evaluation across diverse benchmarks. Combined with AgentLab, a complementary framework that aids in agent creation, testing, and analysis, BrowserGym offers flexibility for integrating new benchmarks while ensuring consistent evaluation and comprehensive experiment management. This standardized approach seeks to reduce the time and complexity of developing web agents, supporting more reliable comparisons and facilitating in-depth analysis of agent behaviors, and could result in more adaptable, capable agents, ultimately accelerating innovation in LLM-driven automation. As a supporting evidence, we conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across all benchmarks currently available in BrowserGym. Among other findings, our results highlight a large discrepancy between OpenAI and Anthropic's latests models, with Claude-3.5-Sonnet leading the way on almost all benchmarks, except on vision-related tasks where GPT-4o is superior. Despite these advancements, our results emphasize that building robust and efficient web agents remains a significant challenge, due to the inherent complexity of real-world web environments and the limitations of current models.
♻ ☆ TorchSISSO: A PyTorch-Based Implementation of the Sure Independence Screening and Sparsifying Operator for Efficient and Interpretable Model Discovery
Symbolic regression (SR) is a powerful machine learning approach that searches for both the structure and parameters of algebraic models, offering interpretable and compact representations of complex data. Unlike traditional regression methods, SR explores progressively complex feature spaces, which can uncover simple models that generalize well, even from small datasets. Among SR algorithms, the Sure Independence Screening and Sparsifying Operator (SISSO) has proven particularly effective in the natural sciences, helping to rediscover fundamental physical laws as well as discover new interpretable equations for materials property modeling. However, its widespread adoption has been limited by performance inefficiencies and the challenges posed by its FORTRAN-based implementation, especially in modern computing environments. In this work, we introduce TorchSISSO, a native Python implementation built in the PyTorch framework. TorchSISSO leverages GPU acceleration, easy integration, and extensibility, offering a significant speed-up and improved accuracy over the original. We demonstrate that TorchSISSO matches or exceeds the performance of the original SISSO across a range of tasks, while dramatically reducing computational time and improving accessibility for broader scientific applications.
♻ ☆ Toward Self-Improvement of LLMs via Imagination, Searching, and Criticizing NeurIPS 2024
Despite the impressive capabilities of Large Language Models (LLMs) on various tasks, they still struggle with scenarios that involves complex reasoning and planning. Recent work proposed advanced prompting techniques and the necessity of fine-tuning with high-quality data to augment LLMs' reasoning abilities. However, these approaches are inherently constrained by data availability and quality. In light of this, self-correction and self-learning emerge as viable solutions, employing strategies that allow LLMs to refine their outputs and learn from self-assessed rewards. Yet, the efficacy of LLMs in self-refining its response, particularly in complex reasoning and planning task, remains dubious. In this paper, we introduce AlphaLLM for the self-improvements of LLMs, which integrates Monte Carlo Tree Search (MCTS) with LLMs to establish a self-improving loop, thereby enhancing the capabilities of LLMs without additional annotations. Drawing inspiration from the success of AlphaGo, AlphaLLM addresses the unique challenges of combining MCTS with LLM for self-improvement, including data scarcity, the vastness search spaces of language tasks, and the subjective nature of feedback in language tasks. AlphaLLM is comprised of prompt synthesis component, an efficient MCTS approach tailored for language tasks, and a trio of critic models for precise feedback. Our experimental results in mathematical reasoning tasks demonstrate that AlphaLLM significantly enhances the performance of LLMs without additional annotations, showing the potential for self-improvement in LLMs.
comment: NeurIPS 2024
♻ ☆ Do graph neural network states contain graph properties?
Deep neural networks (DNNs) achieve state-of-the-art performance on many tasks, but this often requires increasingly larger model sizes, which in turn leads to more complex internal representations. Explainability techniques (XAI) have made remarkable progress in the interpretability of ML models. However, the non-relational nature of Graph neural networks (GNNs) make it difficult to reuse already existing XAI methods. While other works have focused on instance-based explanation methods for GNNs, very few have investigated model-based methods and, to our knowledge, none have tried to probe the embedding of the GNNs for well-known structural graph properties. In this paper we present a model agnostic explainability pipeline for GNNs employing diagnostic classifiers. This pipeline aims to probe and interpret the learned representations in GNNs across various architectures and datasets, refining our understanding and trust in these models.
comment: 10 pages, 22 figures, conference
♻ ☆ ASTD Patterns for Integrated Continuous Anomaly Detection In Data Logs
This paper investigates the use of the ASTD language for ensemble anomaly detection in data logs. It uses a sliding window technique for continuous learning in data streams, coupled with updating learning models upon the completion of each window to maintain accurate detection and align with current data trends. It proposes ASTD patterns for combining learning models, especially in the context of unsupervised learning, which is commonly used for data streams. To facilitate this, a new ASTD operator is proposed, the Quantified Flow, which enables the seamless combination of learning models while ensuring that the specification remains concise. Our contribution is a specification pattern, highlighting the capacity of ASTDs to abstract and modularize anomaly detection systems. The ASTD language provides a unique approach to develop data flow anomaly detection systems, grounded in the combination of processes through the graphical representation of the language operators. This simplifies the design task for developers, who can focus primarily on defining the functional operations that constitute the system.
♻ ☆ Why Fine-grained Labels in Pretraining Benefit Generalization?
Recent studies show that pretraining a deep neural network with fine-grained labeled data, followed by fine-tuning on coarse-labeled data for downstream tasks, often yields better generalization than pretraining with coarse-labeled data. While there is ample empirical evidence supporting this, the theoretical justification remains an open problem. This paper addresses this gap by introducing a "hierarchical multi-view" structure to confine the input data distribution. Under this framework, we prove that: 1) coarse-grained pretraining only allows a neural network to learn the common features well, while 2) fine-grained pretraining helps the network learn the rare features in addition to the common ones, leading to improved accuracy on hard downstream test samples.
comment: arXiv admin note: substantial text overlap with arXiv:2303.16887
♻ ☆ Optimistic Query Routing in Clustering-based Approximate Maximum Inner Product Search
Clustering-based nearest neighbor search is an effective method in which points are partitioned into geometric shards to form an index, with only a few shards searched during query processing to find a set of top-$k$ vectors. Even though the search efficacy is heavily influenced by the algorithm that identifies the shards to probe, it has received little attention in the literature. This work bridges that gap by studying routing in clustering-based maximum inner product search. We unpack existing routers and notice the surprising contribution of optimism. We then take a page from the sequential decision making literature and formalize that insight following the principle of ``optimism in the face of uncertainty.'' In particular, we present a framework that incorporates the moments of the distribution of inner products within each shard to estimate the maximum inner product. We then present an instance of our algorithm that uses only the first two moments to reach the same accuracy as state-of-the-art routers such as ScaNN by probing up to $50\%$ fewer points on benchmark datasets. Our algorithm is also space-efficient: we design a sketch of the second moment whose size is independent of the number of points and requires $\mathcal{O}(1)$ vectors per shard.
♻ ☆ DeMem: Privacy-Enhanced Robust Adversarial Learning via De-Memorization
Adversarial robustness, the ability of a model to withstand manipulated inputs that cause errors, is essential for ensuring the trustworthiness of machine learning models in real-world applications. However, previous studies have shown that enhancing adversarial robustness through adversarial training increases vulnerability to privacy attacks. While differential privacy can mitigate these attacks, it often compromises robustness against both natural and adversarial samples. Our analysis reveals that differential privacy disproportionately impacts low-risk samples, causing an unintended performance drop. To address this, we propose DeMem, which selectively targets high-risk samples, achieving a better balance between privacy protection and model robustness. DeMem is versatile and can be seamlessly integrated into various adversarial training techniques. Extensive evaluations across multiple training methods and datasets demonstrate that DeMem significantly reduces privacy leakage while maintaining robustness against both natural and adversarial samples. These results confirm DeMem's effectiveness and broad applicability in enhancing privacy without compromising robustness.
comment: 10 pages
♻ ☆ AdaSociety: An Adaptive Environment with Social Structures for Multi-Agent Decision-Making NeurIPS
Traditional interactive environments limit agents' intelligence growth with fixed tasks. Recently, single-agent environments address this by generating new tasks based on agent actions, enhancing task diversity. We consider the decision-making problem in multi-agent settings, where tasks are further influenced by social connections, affecting rewards and information access. However, existing multi-agent environments lack a combination of adaptive physical surroundings and social connections, hindering the learning of intelligent behaviors. To address this, we introduce AdaSociety, a customizable multi-agent environment featuring expanding state and action spaces, alongside explicit and alterable social structures. As agents progress, the environment adaptively generates new tasks with social structures for agents to undertake. In AdaSociety, we develop three mini-games showcasing distinct social structures and tasks. Initial results demonstrate that specific social structures can promote both individual and collective benefits, though current reinforcement learning and LLM-based algorithms show limited effectiveness in leveraging social structures to enhance performance. Overall, AdaSociety serves as a valuable research platform for exploring intelligence in diverse physical and social settings. The code is available at https://github.com/bigai-ai/AdaSociety.
comment: Accepted at NeurIPS D&B 2024
♻ ☆ Relaxed Equivariant Graph Neural Networks
3D Euclidean symmetry equivariant neural networks have demonstrated notable success in modeling complex physical systems. We introduce a framework for relaxed $E(3)$ graph equivariant neural networks that can learn and represent symmetry breaking within continuous groups. Building on the existing e3nn framework, we propose the use of relaxed weights to allow for controlled symmetry breaking. We show empirically that these relaxed weights learn the correct amount of symmetry breaking.
comment: Extended abstract presented at the Geometry-grounded Representation Learning and Generative Modeling Workshop (GRaM) at the 41st International Conference on Machine Learning, July 2024, Vienna, Austria
♻ ☆ Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space NeurIPS 2024
Modern generative models demonstrate impressive capabilities, likely stemming from an ability to identify and manipulate abstract concepts underlying their training data. However, fundamental questions remain: what determines the concepts a model learns, the order in which it learns them, and its ability to manipulate those concepts? To address these questions, we propose analyzing a model's learning dynamics via a framework we call the concept space, where each axis represents an independent concept underlying the data generating process. By characterizing learning dynamics in this space, we identify how the speed at which a concept is learned, and hence the order of concept learning, is controlled by properties of the data we term concept signal. Further, we observe moments of sudden turns in the direction of a model's learning dynamics in concept space. Surprisingly, these points precisely correspond to the emergence of hidden capabilities, i.e., where latent interventions show the model possesses the capability to manipulate a concept, but these capabilities cannot yet be elicited via naive input prompting. While our results focus on synthetically defined toy datasets, we hypothesize a general claim on emergence of hidden capabilities may hold: generative models possess latent capabilities that emerge suddenly and consistently during training, though a model might not exhibit these capabilities under naive input prompting.
comment: NeurIPS 2024 (Spotlight)
♻ ☆ Quantum anomaly detection in the latent space of proton collision events at the LHC
The ongoing quest to discover new phenomena at the LHC necessitates the continuous development of algorithms and technologies. Established approaches like machine learning, along with emerging technologies such as quantum computing show promise in the enhancement of experimental capabilities. In this work, we propose a strategy for anomaly detection tasks at the LHC based on unsupervised quantum machine learning, and demonstrate its effectiveness in identifying new phenomena. The designed quantum models, an unsupervised kernel machine and two clustering algorithms, are trained to detect new-physics events using a latent representation of LHC data, generated by an autoencoder designed to accommodate current quantum hardware limitations on problem size. For kernel-based anomaly detection, we implement an instance of the model on a quantum computer, and we identify a regime where it significantly outperforms its classical counterparts. We show that the observed performance enhancement is related to the quantum resources utilised by the model.
comment: Peer-reviewed version, new Figure 4, revised tables, and added some references. 8 Figures, 3 Tables
♻ ☆ AFD: Mitigating Feature Gap for Adversarial Robustness by Feature Disentanglement
Adversarial fine-tuning methods enhance adversarial robustness via fine-tuning the pre-trained model in an adversarial training manner. However, we identify that some specific latent features of adversarial samples are confused by adversarial perturbation and lead to an unexpectedly increasing gap between features in the last hidden layer of natural and adversarial samples. To address this issue, we propose a disentanglement-based approach to explicitly model and further remove the specific latent features. We introduce a feature disentangler to separate out the specific latent features from the features of the adversarial samples, thereby boosting robustness by eliminating the specific latent features. Besides, we align clean features in the pre-trained model with features of adversarial samples in the fine-tuned model, to benefit from the intrinsic features of natural samples. Empirical evaluations on three benchmark datasets demonstrate that our approach surpasses existing adversarial fine-tuning methods and adversarial training baselines.
comment: 7 pages, 5 figures
♻ ☆ Symbolic Regression with a Learned Concept Library NeurIPS
We present a novel method for symbolic regression (SR), the task of searching for compact programmatic hypotheses that best explain a dataset. The problem is commonly solved using genetic algorithms; we show that we can enhance such methods by inducing a library of abstract textual concepts. Our algorithm, called LaSR, uses zero-shot queries to a large language model (LLM) to discover and evolve concepts occurring in known high-performing hypotheses. We discover new hypotheses using a mix of standard evolutionary steps and LLM-guided steps (obtained through zero-shot LLM queries) conditioned on discovered concepts. Once discovered, hypotheses are used in a new round of concept abstraction and evolution. We validate LaSR on the Feynman equations, a popular SR benchmark, as well as a set of synthetic tasks. On these benchmarks, LaSR substantially outperforms a variety of state-of-the-art SR approaches based on deep learning and evolutionary algorithms. Moreover, we show that LaSR can be used to discover a novel and powerful scaling law for LLMs.
comment: NeurIPS version; 10 pages; no checklist; added more experiment details
♻ ☆ Leveraging Large Language Models for Node Generation in Few-Shot Learning on Text-Attributed Graphs AAAI2025
Text-attributed graphs have recently garnered significant attention due to their wide range of applications in web domains. Existing methodologies employ word embedding models for acquiring text representations as node features, which are subsequently fed into Graph Neural Networks (GNNs) for training. Recently, the advent of Large Language Models (LLMs) has introduced their powerful capabilities in information retrieval and text generation, which can greatly enhance the text attributes of graph data. Furthermore, the acquisition and labeling of extensive datasets are both costly and time-consuming endeavors. Consequently, few-shot learning has emerged as a crucial problem in the context of graph learning tasks. In order to tackle this challenge, we propose a lightweight paradigm called LLM4NG, which adopts a plug-and-play approach to empower text-attributed graphs through node generation using LLMs. Specifically, we utilize LLMs to extract semantic information from the labels and generate samples that belong to these categories as exemplars. Subsequently, we employ an edge predictor to capture the structural information inherent in the raw dataset and integrate the newly generated samples into the original graph. This approach harnesses LLMs for enhancing class-level information and seamlessly introduces labeled nodes and edges without modifying the raw dataset, thereby facilitating the node classification task in few-shot scenarios. Extensive experiments demonstrate the outstanding performance of our proposed paradigm, particularly in low-shot scenarios. For instance, in the 1-shot setting of the ogbn-arxiv dataset, LLM4NG achieves a 76% improvement over the baseline model.
comment: Accepted by AAAI2025
♻ ☆ From Biased to Unbiased Dynamics: An Infinitesimal Generator Approach
We investigate learning the eigenfunctions of evolution operators for time-reversal invariant stochastic processes, a prime example being the Langevin equation used in molecular dynamics. Many physical or chemical processes described by this equation involve transitions between metastable states separated by high potential barriers that can hardly be crossed during a simulation. To overcome this bottleneck, data are collected via biased simulations that explore the state space more rapidly. We propose a framework for learning from biased simulations rooted in the infinitesimal generator of the process and the associated resolvent operator. We contrast our approach to more common ones based on the transfer operator, showing that it can provably learn the spectral properties of the unbiased system from biased data. In experiments, we highlight the advantages of our method over transfer operator approaches and recent developments based on generator learning, demonstrating its effectiveness in estimating eigenfunctions and eigenvalues. Importantly, we show that even with datasets containing only a few relevant transitions due to sub-optimal biasing, our approach recovers relevant information about the transition mechanism.
♻ ☆ Which LLMs are Difficult to Detect? A Detailed Analysis of Potential Factors Contributing to Difficulties in LLM Text Detection NeurIPS 2024
As LLMs increase in accessibility, LLM-generated texts have proliferated across several fields, such as scientific, academic, and creative writing. However, LLMs are not created equally; they may have different architectures and training datasets. Thus, some LLMs may be more challenging to detect than others. Using two datasets spanning four total writing domains, we train AI-generated (AIG) text classifiers using the LibAUC library - a deep learning library for training classifiers with imbalanced datasets. Our results in the Deepfake Text dataset show that AIG-text detection varies across domains, with scientific writing being relatively challenging. In the Rewritten Ivy Panda (RIP) dataset focusing on student essays, we find that the OpenAI family of LLMs was substantially difficult for our classifiers to distinguish from human texts. Additionally, we explore possible factors that could explain the difficulties in detecting OpenAI-generated texts.
comment: Accepted at NeurIPS 2024 - Safe Generative AI Workshop; Camera-ready version
♻ ☆ SpaFL: Communication-Efficient Federated Learning with Sparse Models and Low computational Overhead NeurIPS 2024
The large communication and computation overhead of federated learning (FL) is one of the main challenges facing its practical deployment over resource-constrained clients and systems. In this work, SpaFL: a communication-efficient FL framework is proposed to optimize sparse model structures with low computational overhead. In SpaFL, a trainable threshold is defined for each filter/neuron to prune its all connected parameters, thereby leading to structured sparsity. To optimize the pruning process itself, only thresholds are communicated between a server and clients instead of parameters, thereby learning how to prune. Further, global thresholds are used to update model parameters by extracting aggregated parameter importance. The generalization bound of SpaFL is also derived, thereby proving key insights on the relation between sparsity and performance. Experimental results show that SpaFL improves accuracy while requiring much less communication and computing resources compared to sparse baselines. The code is available at https://github.com/news-vt/SpaFL_NeruIPS_2024
comment: Published in NeurIPS 2024
♻ ☆ Unsupervised Learning of Unbiased Visual Representations
Deep neural networks often struggle to learn robust representations in the presence of dataset biases, leading to suboptimal generalization on unbiased datasets. This limitation arises because the models heavily depend on peripheral and confounding factors, inadvertently acquired during training. Existing approaches to address this problem typically involve explicit supervision of bias attributes or reliance on prior knowledge about the biases. In this study, we address the challenging scenario where no explicit annotations of bias are available, and there's no prior knowledge about its nature. We present a fully unsupervised debiasing framework with three key steps: firstly, leveraging the inherent tendency to learn malignant biases to acquire a bias-capturing model; next, employing a pseudo-labeling process to obtain bias labels; and finally, applying cutting-edge supervised debiasing techniques to achieve an unbiased model. Additionally, we introduce a theoretical framework for evaluating model biasedness and conduct a detailed analysis of how biases impact neural network training. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of our method, showcasing state-of-the-art performance in various settings, occasionally surpassing fully supervised debiasing approaches.
comment: Accepted at IEEE Transactions on Artificial Intelligence (TAI)
♻ ☆ Electrocardiogram (ECG) Based Cardiac Arrhythmia Detection and Classification using Machine Learning Algorithms
The rapid advancements in Artificial Intelligence, specifically Machine Learning (ML) and Deep Learning (DL), have opened new prospects in medical sciences for improved diagnosis, prognosis, and treatment of severe health conditions. This paper focuses on the development of an ML model with high predictive accuracy to classify arrhythmic electrocardiogram (ECG) signals. The ECG signals datasets utilized in this study were sourced from the PhysioNet and MIT-BIH databases. The research commenced with binary classification, where an optimized Bidirectional Long Short-Term Memory (Bi-LSTM) model yielded excellent results in differentiating normal and atrial fibrillation signals. A pivotal aspect of this research was a survey among medical professionals, which not only validated the practicality of AI-based ECG classifiers but also identified areas for improvement, including accuracy and the inclusion of more arrhythmia types. These insights drove the development of an advanced Convolutional Neural Network (CNN) system capable of classifying five different types of ECG signals with better accuracy and precision. The CNN model's robust performance was ensured through rigorous stratified 5-fold cross validation. A web portal was also developed to demonstrate real-world utility, offering access to the trained model for real-time classification. This study highlights the potential applications of such models in remote health monitoring, predictive healthcare, assistive diagnostic tools, and simulated environments for educational training and interdisciplinary collaboration between data scientists and medical personnel.
♻ ☆ DeCLIP: Decoding CLIP representations for deepfake localization WACV
Generative models can create entirely new images, but they can also partially modify real images in ways that are undetectable to the human eye. In this paper, we address the challenge of automatically detecting such local manipulations. One of the most pressing problems in deepfake detection remains the ability of models to generalize to different classes of generators. In the case of fully manipulated images, representations extracted from large self-supervised models (such as CLIP) provide a promising direction towards more robust detectors. Here, we introduce DeCLIP, a first attempt to leverage such large pretrained features for detecting local manipulations. We show that, when combined with a reasonably large convolutional decoder, pretrained self-supervised representations are able to perform localization and improve generalization capabilities over existing methods. Unlike previous work, our approach is able to perform localization on the challenging case of latent diffusion models, where the entire image is affected by the fingerprint of the generator. Moreover, we observe that this type of data, which combines local semantic information with a global fingerprint, provides more stable generalization than other categories of generative methods.
comment: Accepted at Winter Conference on Applications of Computer Vision (WACV) 2025
♻ ☆ FreqRISE: Explaining time series using frequency masking
Time-series data are fundamentally important for many critical domains such as healthcare, finance, and climate, where explainable models are necessary for safe automated decision making. To develop explainable artificial intelligence in these domains therefore implies explaining salient information in the time series. Current methods for obtaining saliency maps assume localized information in the raw input space. In this paper, we argue that the salient information of a number of time series is more likely to be localized in the frequency domain. We propose FreqRISE, which uses masking-based methods to produce explanations in the frequency and time-frequency domain, and outperforms strong baselines across a number of tasks. The source code is available here: \url{https://github.com/theabrusch/FreqRISE}.
comment: Accepted at the Northern Lights Deep Learning Conference 2025
♻ ☆ Bigger is not Always Better: Scaling Properties of Latent Diffusion Models
We study the scaling properties of latent diffusion models (LDMs) with an emphasis on their sampling efficiency. While improved network architecture and inference algorithms have shown to effectively boost sampling efficiency of diffusion models, the role of model size -- a critical determinant of sampling efficiency -- has not been thoroughly examined. Through empirical analysis of established text-to-image diffusion models, we conduct an in-depth investigation into how model size influences sampling efficiency across varying sampling steps. Our findings unveil a surprising trend: when operating under a given inference budget, smaller models frequently outperform their larger equivalents in generating high-quality results. Moreover, we extend our study to demonstrate the generalizability of the these findings by applying various diffusion samplers, exploring diverse downstream tasks, evaluating post-distilled models, as well as comparing performance relative to training compute. These findings open up new pathways for the development of LDM scaling strategies which can be employed to enhance generative capabilities within limited inference budgets.
comment: Accepted to TMLR. Camera-ready version
♻ ☆ RectifID: Personalizing Rectified Flow with Anchored Classifier Guidance NeurIPS 2024
Customizing diffusion models to generate identity-preserving images from user-provided reference images is an intriguing new problem. The prevalent approaches typically require training on extensive domain-specific images to achieve identity preservation, which lacks flexibility across different use cases. To address this issue, we exploit classifier guidance, a training-free technique that steers diffusion models using an existing classifier, for personalized image generation. Our study shows that based on a recent rectified flow framework, the major limitation of vanilla classifier guidance in requiring a special classifier can be resolved with a simple fixed-point solution, allowing flexible personalization with off-the-shelf image discriminators. Moreover, its solving procedure proves to be stable when anchored to a reference flow trajectory, with a convergence guarantee. The derived method is implemented on rectified flow with different off-the-shelf image discriminators, delivering advantageous personalization results for human faces, live subjects, and certain objects. Code is available at https://github.com/feifeiobama/RectifID.
comment: NeurIPS 2024
♻ ☆ Guiding Vision-Language Model Selection for Visual Question-Answering Across Tasks, Domains, and Knowledge Types
Visual Question-Answering (VQA) has become key to user experience, particularly after improved generalization capabilities of Vision-Language Models (VLMs). But evaluating VLMs for an application requirement using a standardized framework in practical settings is still challenging. This paper aims to solve that using an end-to-end framework. We present VQA360 - a novel dataset derived from established VQA benchmarks, annotated with task types, application domains, and knowledge types, for a comprehensive evaluation. We also introduce GoEval, a multimodal evaluation metric developed using GPT-4o, achieving a correlation factor of 56.71% with human judgments. Our experiments with state-of-the-art VLMs reveal that no single model excels universally, thus, making a right choice a key design decision. Proprietary models such as Gemini-1.5-Pro and GPT-4o-mini generally outperform others, but open-source models like InternVL-2-8B and CogVLM-2-Llama-3-19B also demonstrate competitive strengths, while providing additional advantages. Our framework can also be extended to other tasks.
comment: 8 pages + references + 6 pages of Appendix
♻ ☆ Guided Quantum Compression for High Dimensional Data Classification
Quantum machine learning provides a fundamentally different approach to analyzing data. However, many interesting datasets are too complex for currently available quantum computers. Present quantum machine learning applications usually diminish this complexity by reducing the dimensionality of the data, e.g., via auto-encoders, before passing it through the quantum models. Here, we design a classical-quantum paradigm that unifies the dimensionality reduction task with a quantum classification model into a single architecture: the guided quantum compression model. We exemplify how this architecture outperforms conventional quantum machine learning approaches on a challenging binary classification problem: identifying the Higgs boson in proton-proton collisions at the LHC. Furthermore, the guided quantum compression model shows better performance compared to the deep learning benchmark when using solely the kinematic variables in our dataset.
comment: Peer-reviewed version, 9 pages, 3 figures
♻ ☆ Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Change KDD
Counterfactual explanations (CFEs) guide users on how to adjust inputs to machine learning models to achieve desired outputs. While existing research primarily addresses static scenarios, real-world applications often involve data or model changes, potentially invalidating previously generated CFEs and rendering user-induced input changes ineffective. Current methods addressing this issue often support only specific models or change types, require extensive hyperparameter tuning, or fail to provide probabilistic guarantees on CFE robustness to model changes. This paper proposes a novel approach for generating CFEs that provides probabilistic guarantees for any model and change type, while offering interpretable and easy-to-select hyperparameters. We establish a theoretical framework for probabilistically defining robustness to model change and demonstrate how our BetaRCE method directly stems from it. BetaRCE is a post-hoc method applied alongside a chosen base CFE generation method to enhance the quality of the explanation beyond robustness. It facilitates a transition from the base explanation to a more robust one with user-adjusted probability bounds. Through experimental comparisons with baselines, we show that BetaRCE yields robust, most plausible, and closest to baseline counterfactual explanations.
comment: Accepted at 31st SIGKDD Conference on Knowledge Discovery and Data Mining 2025
♻ ☆ Number Theoretic Accelerated Learning of Physics-Informed Neural Networks
Physics-informed neural networks solve partial differential equations by training neural networks. Since this method approximates infinite-dimensional PDE solutions with finite collocation points, minimizing discretization errors by selecting suitable points is essential for accelerating the learning process. Inspired by number theoretic methods for numerical analysis, we introduce good lattice training and periodization tricks, which ensure the conditions required by the theory. Our experiments demonstrate that GLT requires 2-7 times fewer collocation points, resulting in lower computational cost, while achieving competitive performance compared to typical sampling methods.
♻ ☆ Score-Based Multimodal Autoencoder
Multimodal Variational Autoencoders (VAEs) represent a promising group of generative models that facilitate the construction of a tractable posterior within the latent space given multiple modalities. Previous studies have shown that as the number of modalities increases, the generative quality of each modality declines. In this study, we explore an alternative approach to enhance the generative performance of multimodal VAEs by jointly modeling the latent space of independently trained unimodal VAEs using score-based models (SBMs). The role of the SBM is to enforce multimodal coherence by learning the correlation among the latent variables. Consequently, our model combines a better generative quality of unimodal VAEs with coherent integration across different modalities using the latent score-based model. In addition, our approach provides the best unconditional coherence.
♻ ☆ AutoDetect: Towards a Unified Framework for Automated Weakness Detection in Large Language Models EMNLP 2024
Although Large Language Models (LLMs) are becoming increasingly powerful, they still exhibit significant but subtle weaknesses, such as mistakes in instruction-following or coding tasks. As these unexpected errors could lead to severe consequences in practical deployments, it is crucial to investigate the limitations within LLMs systematically. Traditional benchmarking approaches cannot thoroughly pinpoint specific model deficiencies, while manual inspections are costly and not scalable. In this paper, we introduce a unified framework, AutoDetect, to automatically expose weaknesses in LLMs across various tasks. Inspired by the educational assessment process that measures students' learning outcomes, AutoDetect consists of three LLM-powered agents: Examiner, Questioner, and Assessor. The collaboration among these three agents is designed to realize comprehensive and in-depth weakness identification. Our framework demonstrates significant success in uncovering flaws, with an identification success rate exceeding 30% in prominent models such as ChatGPT and Claude. More importantly, these identified weaknesses can guide specific model improvements, proving more effective than untargeted data augmentation methods like Self-Instruct. Our approach has led to substantial enhancements in popular LLMs, including the Llama series and Mistral-7b, boosting their performance by over 10% across several benchmarks. Code and data are publicly available at https://github.com/thu-coai/AutoDetect.
comment: EMNLP 2024 findings
♻ ☆ Divergent Ensemble Networks: Enhancing Uncertainty Estimation with Shared Representations and Independent Branching
Ensemble learning has proven effective in improving predictive performance and estimating uncertainty in neural networks. However, conventional ensemble methods often suffer from redundant parameter usage and computational inefficiencies due to entirely independent network training. To address these challenges, we propose the Divergent Ensemble Network (DEN), a novel architecture that combines shared representation learning with independent branching. DEN employs a shared input layer to capture common features across all branches, followed by divergent, independently trainable layers that form an ensemble. This shared-to-branching structure reduces parameter redundancy while maintaining ensemble diversity, enabling efficient and scalable learning.
♻ ☆ Contrasting Adversarial Perturbations: The Space of Harmless Perturbations
Existing works have extensively studied adversarial examples, which are minimal perturbations that can mislead the output of deep neural networks (DNNs) while remaining imperceptible to humans. However, in this work, we reveal the existence of a harmless perturbation space, in which perturbations drawn from this space, regardless of their magnitudes, leave the network output unchanged when applied to inputs. Essentially, the harmless perturbation space emerges from the usage of non-injective functions (linear or non-linear layers) within DNNs, enabling multiple distinct inputs to be mapped to the same output. For linear layers with input dimensions exceeding output dimensions, any linear combination of the orthogonal bases of the nullspace of the parameter consistently yields no change in their output. For non-linear layers, the harmless perturbation space may expand, depending on the properties of the layers and input samples. Inspired by this property of DNNs, we solve for a family of general perturbation spaces that are redundant for the DNN's decision, and can be used to hide sensitive data and serve as a means of model identification. Our work highlights the distinctive robustness of DNNs (i.e., consistency under large magnitude perturbations) in contrast to adversarial examples (vulnerability for small imperceptible noises).
♻ ☆ Non-Parametric Learning of Stochastic Differential Equations with Non-asymptotic Fast Rates of Convergence
We propose a novel non-parametric learning paradigm for the identification of drift and diffusion coefficients of multi-dimensional non-linear stochastic differential equations, which relies upon discrete-time observations of the state. The key idea essentially consists of fitting a RKHS-based approximation of the corresponding Fokker-Planck equation to such observations, yielding theoretical estimates of non-asymptotic learning rates which, unlike previous works, become increasingly tighter when the regularity of the unknown drift and diffusion coefficients becomes higher. Our method being kernel-based, offline pre-processing may be profitably leveraged to enable efficient numerical implementation, offering excellent balance between precision and computational complexity.
♻ ☆ A Comprehensive Study of Shapley Value in Data Analytics
Over the recent years, Shapley value (SV), a solution concept from cooperative game theory, has found numerous applications in data analytics (DA). This paper provides the first comprehensive study of SV used throughout the DA workflow, which involves three main steps: data fabric, data exploration, and result reporting. We summarize existing versatile forms of SV used in these steps by a unified definition and clarify the essential functionalities that SV can provide for data scientists. We categorize the arts in this field based on the technical challenges they tackled, which include computation efficiency, approximation error, privacy preservation, and appropriate interpretations. We discuss these challenges and analyze the corresponding solutions. We also implement SVBench, the first open-sourced benchmark for developing SV applications, and conduct experiments on six DA tasks to validate our analysis and discussions. Based on the qualitative and quantitative results, we identify the limitations of current efforts for applying SV to DA and highlight the directions of future research and engineering.
♻ ☆ Separation Power of Equivariant Neural Networks
The separation power of a machine learning model refers to its ability to distinguish between different inputs and is often used as a proxy for its expressivity. Indeed, knowing the separation power of a family of models is a necessary condition to obtain fine-grained universality results. In this paper, we analyze the separation power of equivariant neural networks, such as convolutional and permutation-invariant networks. We first present a complete characterization of inputs indistinguishable by models derived by a given architecture. From this results, we derive how separability is influenced by hyperparameters and architectural choices-such as activation functions, depth, hidden layer width, and representation types. Notably, all non-polynomial activations, including ReLU and sigmoid, are equivalent in expressivity and reach maximum separation power. Depth improves separation power up to a threshold, after which further increases have no effect. Adding invariant features to hidden representations does not impact separation power. Finally, block decomposition of hidden representations affects separability, with minimal components forming a hierarchy in separation power that provides a straightforward method for comparing the separation power of models.
comment: 10 pages of main text, 1 figure
♻ ☆ Enhancing Vision-Language Model Pre-training with Image-text Pair Pruning Based on Word Frequency
We propose Word-Frequency-based Image-Text Pair Pruning (WFPP), a novel data pruning method that improves the efficiency of VLMs. Unlike MetaCLIP, our method does not need metadata for pruning, but selects text-image pairs to prune based on the content of the text. Specifically, WFPP prunes text-image pairs containing high-frequency words across the entire training dataset. The effect of WFPP is to reduce the dominance of frequent words. The result a better balanced word-frequency distribution in the dataset, which is known to improve the training of word embedding models. After pre-training on the pruned subset, we fine-tuned the model on the entire dataset for one additional epoch to achieve better performance. Our experiments demonstrate that applying WFPP when training a CLIP model improves performance on a wide range of downstream tasks. WFPP also provides the advantage of speeding up pre-training by using fewer samples. Additionally, we analyze the training data before and after pruning to visualize how WFPP changes the balance of word frequencies. We hope our work encourages researchers to consider the distribution of words in the training data when pre-training VLMs, not limited to CLIP.
♻ ☆ XAMI -- A Benchmark Dataset for Artefact Detection in XMM-Newton Optical Images SP
Reflected or scattered light produce artefacts in astronomical observations that can negatively impact the scientific study. Hence, automated detection of these artefacts is highly beneficial, especially with the increasing amounts of data gathered. Machine learning methods are well-suited to this problem, but currently there is a lack of annotated data to train such approaches to detect artefacts in astronomical observations. In this work, we present a dataset of images from the XMM-Newton space telescope Optical Monitoring camera showing different types of artefacts. We hand-annotated a sample of 1000 images with artefacts which we use to train automated ML methods. We further demonstrate techniques tailored for accurate detection and masking of artefacts using instance segmentation. We adopt a hybrid approach, combining knowledge from both convolutional neural networks (CNNs) and transformer-based models and use their advantages in segmentation. The presented method and dataset will advance artefact detection in astronomical observations by providing a reproducible baseline. All code and data are made available (https://github.com/ESA-Datalabs/XAMI-model and https://github.com/ESA-Datalabs/XAMI-dataset).
comment: Accepted for oral presentation at SPAICE 2024
♻ ☆ Probability of Collision of satellites and space debris for short-term encounters: Rederivation and fast-to-compute upper and lower bounds
The proliferation of space debris in LEO has become a major concern for the space industry. With the growing interest in space exploration, the prediction of potential collisions between objects in orbit has become a crucial issue. It is estimated that, in orbit, there are millions of fragments a few millimeters in size and thousands of inoperative satellites and discarded rocket stages. Given the high speeds that these fragments can reach, even fragments a few millimeters in size can cause fractures in a satellite's hull or put a serious crack in the window of a space shuttle. The conventional method proposed by Akella and Alfriend in 2000 remains widely used to estimate the probability of collision in short-term encounters. Given the small period of time, it is assumed that, during the encounter: (1) trajectories are represented by straight lines with constant velocity; (2) there is no velocity uncertainty and the position exhibits a stationary distribution throughout the encounter; and (3) position uncertainties are independent and represented by Gaussian distributions. This study introduces a novel derivation based on first principles that naturally allows for tight and fast upper and lower bounds for the probability of collision. We tested implementations of both probability and bound computations with the original and our formulation on a real CDM dataset used in ESA's Collision Avoidance Challenge. Our approach reduces the calculation of the probability to two one-dimensional integrals and has the potential to significantly reduce the processing time compared to the traditional method, from 80% to nearly real-time.
♻ ☆ MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control
Autonomous agents powered by large language models (LLMs) show promising potential in assistive tasks across various domains, including mobile device control. As these agents interact directly with personal information and device settings, ensuring their safe and reliable behavior is crucial to prevent undesirable outcomes. However, no benchmark exists for standardized evaluation of the safety of mobile device-control agents. In this work, we introduce MobileSafetyBench, a benchmark designed to evaluate the safety of device-control agents within a realistic mobile environment based on Android emulators. We develop a diverse set of tasks involving interactions with various mobile applications, including messaging and banking applications, challenging agents with managing risks encompassing misuse and negative side effects. These tasks include tests to evaluate the safety of agents in daily scenarios as well as their robustness against indirect prompt injection attacks. Our experiments demonstrate that baseline agents, based on state-of-the-art LLMs, often fail to effectively prevent harm while performing the tasks. To mitigate these safety concerns, we propose a prompting method that encourages agents to prioritize safety considerations. While this method shows promise in promoting safer behaviors, there is still considerable room for improvement to fully earn user trust. This highlights the urgent need for continued research to develop more robust safety mechanisms in mobile environments. We open-source our benchmark at: https://mobilesafetybench.github.io/.
♻ ☆ Using Large Language Models for Expert Prior Elicitation in Predictive Modelling
Large language models (LLMs), trained on diverse data effectively acquire a breadth of information across various domains. However, their computational complexity, cost, and lack of transparency hinder their direct application for specialised tasks. In fields such as clinical research, acquiring expert annotations or prior knowledge about predictive models is often costly and time-consuming. This study proposes the use of LLMs to elicit expert prior distributions for predictive models. This approach also provides an alternative to in-context learning, where language models are tasked with making predictions directly. In this work, we compare LLM-elicited and uninformative priors, evaluate whether LLMs truthfully generate parameter distributions, and propose a model selection strategy for in-context learning and prior elicitation. Our findings show that LLM-elicited prior parameter distributions significantly reduce predictive error compared to uninformative priors in low-data settings. Applied to clinical problems, this translates to fewer required biological samples, lowering cost and resources. Prior elicitation also consistently outperforms and proves more reliable than in-context learning at a lower cost, making it a preferred alternative in our setting. We demonstrate the utility of this method across various use cases, including clinical applications. For infection prediction, using LLM-elicited priors reduced the number of required labels to achieve the same accuracy as an uninformative prior by 55%, 200 days earlier in the study.
♻ ☆ pfl-research: simulation framework for accelerating research in Private Federated Learning
Federated learning (FL) is an emerging machine learning (ML) training paradigm where clients own their data and collaborate to train a global model, without revealing any data to the server and other participants. Researchers commonly perform experiments in a simulation environment to quickly iterate on ideas. However, existing open-source tools do not offer the efficiency required to simulate FL on larger and more realistic FL datasets. We introduce pfl-research, a fast, modular, and easy-to-use Python framework for simulating FL. It supports TensorFlow, PyTorch, and non-neural network models, and is tightly integrated with state-of-the-art privacy algorithms. We study the speed of open-source FL frameworks and show that pfl-research is 7-72$\times$ faster than alternative open-source frameworks on common cross-device setups. Such speedup will significantly boost the productivity of the FL research community and enable testing hypotheses on realistic FL datasets that were previously too resource intensive. We release a suite of benchmarks that evaluates an algorithm's overall performance on a diverse set of realistic scenarios. The code is available on GitHub at https://github.com/apple/pfl-research.
♻ ☆ Approximating Langevin Monte Carlo with ResNet-like Neural Network architectures
We sample from a given target distribution by constructing a neural network which maps samples from a simple reference, e.g. the standard normal distribution, to samples from the target. To that end, we propose using a neural network architecture inspired by the Langevin Monte Carlo (LMC) algorithm. Based on LMC perturbation results, we show approximation rates of the proposed architecture for smooth, log-concave target distributions measured in the Wasserstein-$2$ distance. The analysis heavily relies on the notion of sub-Gaussianity of the intermediate measures of the perturbed LMC process. In particular, we derive bounds on the growth of the intermediate variance proxies under different assumptions on the perturbations. Moreover, we propose an architecture similar to deep residual neural networks and derive expressivity results for approximating the sample to target distribution map.
♻ ☆ An inferential measure of dependence between two systems using Bayesian model comparison
We propose to quantify dependence between two systems $X$ and $Y$ in a dataset $D$ based on the Bayesian comparison of two models: one, $H_0$, of statistical independence and another one, $H_1$, of dependence. In this framework, dependence between $X$ and $Y$ in $D$, denoted $B(X,Y|D)$, is quantified as $P(H_1|D)$, the posterior probability for the model of dependence given $D$, or any strictly increasing function thereof. It is therefore a measure of the evidence for dependence between $X$ and $Y$ as modeled by $H_1$ and observed in $D$. We review several statistical models and reconsider standard results in the light of $B(X,Y|D)$ as a measure of dependence. Using simulations, we focus on two specific issues: the effect of noise and the behavior of $B(X,Y|D)$ when $H_1$ has a parameter coding for the intensity of dependence. We then derive some general properties of $B(X,Y|D)$, showing that it quantifies the information contained in $D$ in favor of $H_1$ versus $H_0$. While some of these properties are typical of what is expected from a valid measure of dependence, others are novel and naturally appear as desired features for specific measures of dependence, which we call inferential. We finally put these results in perspective; in particular, we discuss the consequences of using the Bayesian framework as well as the similarities and differences between $B(X,Y|D)$ and mutual information.
comment: To be published in IEEE Transaction on Systems, Man, and Cybernetics: Systems
♻ ☆ A Comprehensive Survey on Kolmogorov Arnold Networks (KAN)
Through this comprehensive survey of Kolmogorov-Arnold Networks(KAN), we have gained a thorough understanding of its theoretical foundation, architectural design, application scenarios, and current research progress. KAN, with its unique architecture and flexible activation functions, excels in handling complex data patterns and nonlinear relationships, demonstrating wide-ranging application potential. While challenges remain, KAN is poised to pave the way for innovative solutions in various fields, potentially revolutionizing how we approach complex computational problems.
♻ ☆ CSGO: Constrained-Softassign Gradient Optimization For Large Graph Matching
Graph matching aims to find correspondences between two graphs. This paper integrates several well-known graph matching algorithms into a framework: the constrained gradient method. The primary difference among these algorithms lies in tuning a step size parameter and constraining operators. By leveraging these insights, we propose an adaptive step size parameter to guarantee the underlying algorithms' convergence, simultaneously enhancing their efficiency and robustness. For the constraining operator, we introduce a scalable softassign for large graph matching problems. Compared to the original softassign, our approach offers increased speed, improved robustness, and reduced risk of overflow. The advanced constraining operator enables a CSGO for large graph matching, which outperforms state-of-the-art methods in experiments. Notably, in attributed graph matching tasks, CSGO achieves an over 10X increase in speed compared to current constrained gradient algorithms.
♻ ☆ RST-LoRA: A Discourse-Aware Low-Rank Adaptation for Long Document Abstractive Summarization NAACL 2024
For long document summarization, discourse structure is important to discern the key content of the text and the differences in importance level between sentences. Unfortunately, the integration of rhetorical structure theory (RST) into parameter-efficient fine-tuning strategies for long document summarization remains unexplored. Therefore, this paper introduces RST-LoRA and proposes four RST-aware variants to explicitly incorporate RST into the LoRA model. Our empirical evaluation demonstrates that incorporating the type and uncertainty of rhetorical relations can complementarily enhance the performance of LoRA in summarization tasks. Furthermore, the best-performing variant we introduced outperforms the vanilla LoRA and full-parameter fine-tuning models, as confirmed by multiple automatic and human evaluations, and even surpasses previous state-of-the-art methods.
comment: NAACL 2024 Main & Long Conference Paper (Oral Presentation)
♻ ☆ Using Machine Bias To Measure Human Bias
Biased human decisions have consequential impacts across various domains, yielding unfair treatment of individuals and resulting in suboptimal outcomes for organizations and society. In recognition of this fact, organizations regularly design and deploy interventions aimed at mitigating these biases. However, measuring human decision biases remains an important but elusive task. Organizations are frequently concerned with mistaken decisions disproportionately affecting one group. In practice, however, this is typically not possible to assess due to the scarcity of a gold standard: a label that indicates what the correct decision would have been. In this work, we propose a machine learning-based framework to assess bias in human-generated decisions when gold standard labels are scarce. We provide theoretical guarantees and empirical evidence demonstrating the superiority of our method over existing alternatives. This proposed methodology establishes a foundation for transparency in human decision-making, carrying substantial implications for managerial duties, and offering potential for alleviating algorithmic biases when human decisions are used as labels to train algorithms.
♻ ☆ SciNews: From Scholarly Complexities to Public Narratives -- A Dataset for Scientific News Report Generation LREC
Scientific news reports serve as a bridge, adeptly translating complex research articles into reports that resonate with the broader public. The automated generation of such narratives enhances the accessibility of scholarly insights. In this paper, we present a new corpus to facilitate this paradigm development. Our corpus comprises a parallel compilation of academic publications and their corresponding scientific news reports across nine disciplines. To demonstrate the utility and reliability of our dataset, we conduct an extensive analysis, highlighting the divergences in readability and brevity between scientific news narratives and academic manuscripts. We benchmark our dataset employing state-of-the-art text generation models. The evaluation process involves both automatic and human evaluation, which lays the groundwork for future explorations into the automated generation of scientific news reports. The dataset and code related to this work are available at https://dongqi.me/projects/SciNews.
comment: LREC-COLING 2024 Main Conference Paper
♻ ☆ ChatGPT vs Human-authored Text: Insights into Controllable Text Summarization and Sentence Style Transfer ACL
Large-scale language models, like ChatGPT, have garnered significant media attention and stunned the public with their remarkable capacity for generating coherent text from short natural language prompts. In this paper, we aim to conduct a systematic inspection of ChatGPT's performance in two controllable generation tasks, with respect to ChatGPT's ability to adapt its output to different target audiences (expert vs. layman) and writing styles (formal vs. informal). Additionally, we evaluate the faithfulness of the generated text, and compare the model's performance with human-authored texts. Our findings indicate that the stylistic variations produced by humans are considerably larger than those demonstrated by ChatGPT, and the generated texts diverge from human samples in several characteristics, such as the distribution of word types. Moreover, we observe that ChatGPT sometimes incorporates factual errors or hallucinations when adapting the text to suit a specific style.
comment: ACL-SRW 2023
♻ ☆ Incorporating Distributions of Discourse Structure for Long Document Abstractive Summarization ACL 2023
For text summarization, the role of discourse structure is pivotal in discerning the core content of a text. Regrettably, prior studies on incorporating Rhetorical Structure Theory (RST) into transformer-based summarization models only consider the nuclearity annotation, thereby overlooking the variety of discourse relation types. This paper introduces the 'RSTformer', a novel summarization model that comprehensively incorporates both the types and uncertainty of rhetorical relations. Our RST-attention mechanism, rooted in document-level rhetorical structure, is an extension of the recently devised Longformer framework. Through rigorous evaluation, the model proposed herein exhibits significant superiority over state-of-the-art models, as evidenced by its notable performance on several automatic metrics and human evaluation.
comment: ACL 2023 (Main conference)
♻ ☆ Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision NeurIPS 2024
Current AI alignment methodologies rely on human-provided demonstrations or judgments, and the learned capabilities of AI systems would be upper-bounded by human capabilities as a result. This raises a challenging research question: How can we keep improving the systems when their capabilities have surpassed the levels of humans? This paper answers this question in the context of tackling hard reasoning tasks (e.g., level 4-5 MATH problems) via learning from human annotations on easier tasks (e.g., level 1-3 MATH problems), which we term as easy-to-hard generalization. Our key insight is that an evaluator (reward model) trained on supervisions for easier tasks can be effectively used for scoring candidate solutions of harder tasks and hence facilitating easy-to-hard generalization over different levels of tasks. Based on this insight, we propose a novel approach to scalable alignment, which firstly trains the (process-supervised) reward models on easy problems (e.g., level 1-3), and then uses them to evaluate the performance of policy models on hard problems. We show that such easy-to-hard generalization from evaluators can enable easy-to-hard generalizations in generators either through re-ranking or reinforcement learning (RL). Notably, our process-supervised 7b RL model and 34b model (reranking@1024) achieves an accuracy of 34.0% and 52.5% on MATH500, respectively, despite only using human supervision on easy problems. Our approach suggests a promising path toward AI systems that advance beyond the frontier of human supervision.
comment: Accepted at NeurIPS 2024
♻ ☆ Resource and Mobility Management in Hybrid LiFi and WiFi Networks: A User-Centric Learning Approach
Hybrid light fidelity (LiFi) and wireless fidelity (WiFi) networks (HLWNets) are an emerging indoor wireless communication paradigm, which combines the advantages of the capacious optical spectra of LiFi and ubiquitous coverage of WiFi. Meanwhile, load balancing (LB) becomes a key challenge in resource management for such hybrid networks. The existing LB methods are mostly network-centric, relying on a central unit to make a solution for the users all at once. Consequently, the solution needs to be updated for all users at the same pace, regardless of their moving status. This would affect the network performance in two aspects: i) when the update frequency is low, it would compromise the connectivity of fast-moving users; ii) when the update frequency is high, it would cause unnecessary handovers as well as hefty feedback costs for slow-moving users. Motivated by this, we investigate user-centric LB which allows users to update their solutions at different paces. The research is developed upon our previous work on adaptive target-condition neural network (ATCNN), which can conduct LB for individual users in quasi-static channels. In this paper, a deep neural network (DNN) model is designed to enable an adaptive update interval for each individual user. This new model is termed as mobility-supporting neural network (MSNN). Associating MSNN with ATCNN, a user-centric LB framework named mobility-supporting ATCNN (MS-ATCNN) is proposed to handle resource management and mobility management simultaneously. Results show that at the same level of average update interval, MS-ATCNN can achieve a network throughput up to 215\% higher than conventional LB methods such as game theory, especially for a larger number of users. In addition, MS-ATCNN costs an ultra low runtime at the level of 100s $\mu$s, which is two to three orders of magnitude lower than game theory.
comment: 13 pages, 13 figures, 4 tables, accepted by IEEE Transactions on Wireless Communications
♻ ☆ Multigraph Message Passing with Bi-Directional Multi-Edge Aggregations
Graph Neural Networks (GNNs) have seen significant advances in recent years, yet their application to multigraphs, where parallel edges exist between the same pair of nodes, remains under-explored. Standard GNNs, designed for simple graphs, compute node representations by combining all connected edges at once, without distinguishing between edges from different neighbors. There are some GNN architectures proposed specifically for multigraphs, yet these architectures perform only node-level aggregation in their message passing layers, which limits their expressive power. Furthermore, these approaches either lack permutation equivariance when a strict total edge ordering is absent, or fail to preserve the topological structure of the multigraph. To address all these shortcomings, we propose MEGA-GNN, a unified framework for message passing on multigraphs that can effectively perform diverse graph learning tasks. Our approach introduces a two-stage aggregation process in the message passing layers: first, parallel edges are aggregated, followed by a node-level aggregation of messages from distinct neighbors. We show that MEGA-GNN is not only permutation equivariant but also universal given a strict total ordering on the edges. Experiments show that MEGA-GNN significantly outperforms state-of-the-art solutions by up to 13\% on Anti-Money Laundering datasets and is on par with their accuracy on real-world phishing classification datasets in terms of minority class F1 score.
comment: 19 pages, 5 figures
♻ ☆ ProcessBench: Identifying Process Errors in Mathematical Reasoning
As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.
♻ ☆ Track4Gen: Teaching Video Diffusion Models to Track Points Improves Video Generation
While recent foundational video generators produce visually rich output, they still struggle with appearance drift, where objects gradually degrade or change inconsistently across frames, breaking visual coherence. We hypothesize that this is because there is no explicit supervision in terms of spatial tracking at the feature level. We propose Track4Gen, a spatially aware video generator that combines video diffusion loss with point tracking across frames, providing enhanced spatial supervision on the diffusion features. Track4Gen merges the video generation and point tracking tasks into a single network by making minimal changes to existing video generation architectures. Using Stable Video Diffusion as a backbone, Track4Gen demonstrates that it is possible to unify video generation and point tracking, which are typically handled as separate tasks. Our extensive evaluations show that Track4Gen effectively reduces appearance drift, resulting in temporally stable and visually coherent video generation. Project page: hyeonho99.github.io/track4gen
comment: Project page: hyeonho99.github.io/track4gen
♻ ☆ Graph Neural Networks for Quantifying Compatibility Mechanisms in Traditional Chinese Medicine
Traditional Chinese Medicine (TCM) involves complex compatibility mechanisms characterized by multi-component and multi-target interactions, which are challenging to quantify. To address this challenge, we applied graph artificial intelligence to develop a TCM multi-dimensional knowledge graph that bridges traditional TCM theory and modern biomedical science (https://zenodo.org/records/13763953 ). Using feature engineering and embedding, we processed key TCM terminology and Chinese herbal pieces (CHP), introducing medicinal properties as virtual nodes and employing graph neural networks with attention mechanisms to model and analyze 6,080 Chinese herbal formulas (CHF). Our method quantitatively assessed the roles of CHP within CHF and was validated using 215 CHF designed for COVID-19 management. With interpretable models, open-source data, and code (https://github.com/ZENGJingqi/GraphAI-for-TCM ), this study provides robust tools for advancing TCM theory and drug discovery.
comment: 10 pages, 5 figures. Includes open-source dataset and code for reproducibility
♻ ☆ Group & Reweight: A Novel Cost-Sensitive Approach to Mitigating Class Imbalance in Network Traffic Classification
Internet services have led to the eruption of network traffic, and machine learning on these Internet data has become an indispensable tool, especially when the application is risk-sensitive. This paper focuses on network traffic classification in the presence of severe class imbalance. Such a distributional trait mostly drifts the optimal decision boundary and results in an unsatisfactory solution. This raises safety concerns in the network traffic field when previous class imbalance methods hardly deal with numerous minority malicious classes. To alleviate these effects, we design a \textit{group \& reweight} strategy for alleviating class imbalance. Inspired by the group distributionally optimization framework, our approach heuristically clusters classes into groups, iteratively updates the non-parametric weights for separate classes, and optimizes the learning model by minimizing reweighted losses. We theoretically interpret the optimization process from a Stackelberg game and perform extensive experiments on typical benchmarks. Results show that our approach can not only suppress the negative effect of class imbalance but also improve the comprehensive performance in prediction.
comment: 21 pages, 10 figures
♻ ☆ Policy Filtration in RLHF to Fine-Tune LLM for Code Generation
Reinforcement learning from human feedback (RLHF) is one of the key techniques that helps large language models (LLMs) to follow instructions and provide helpful and harmless responses. While direct policy optimization methods exist, state-of-the-art LLMs adopt RL-based methods (usually PPO) in RLHF to train the policy to generate good responses guided by a reward model learned from preference data. The main challenge of these methods is the inaccuracy of the intermediate reward model, especially in code generation tasks that require long and complex reasoning to score a response. We find that the reliability of the reward model varies across responses assigned with different rewards. This motivates us to filter the samples whose rewards may be unreliable to improve signal-to-noise ratio during policy learning, resulting in Policy Filtration for Proximal Policy Optimization (PF-PPO). To choose a proper policy filtration strategy for a given reward model, the coefficient of determination ($R^2$) between rewards and actual scores on filtered samples serves as a good metrics and helps us find several promising strategies. We provide extensive experiments to validate the effectiveness of PF-PPO in code generation tasks, and find that some variants of PF-PPO are highly effective and achieve new state-of-the-art performance across 7-billion-parameter models on HumanEval, MBPP, and a new and more challenging LeetCode Contest benchmark.
♻ ☆ Nonparametric Classification on Low Dimensional Manifolds using Overparameterized Convolutional Residual Networks
Convolutional residual neural networks (ConvResNets), though overparameterized, can achieve remarkable prediction performance in practice, which cannot be well explained by conventional wisdom. To bridge this gap, we study the performance of ConvResNeXts, which cover ConvResNets as a special case, trained with weight decay from the perspective of nonparametric classification. Our analysis allows for infinitely many building blocks in ConvResNeXts, and shows that weight decay implicitly enforces sparsity on these blocks. Specifically, we consider a smooth target function supported on a low-dimensional manifold, then prove that ConvResNeXts can adapt to the function smoothness and low-dimensional structures and efficiently learn the function without suffering from the curse of dimensionality. Our findings partially justify the advantage of overparameterized ConvResNeXts over conventional machine learning models.
comment: 20 pages, 1 figure
♻ ☆ A Resolution Independent Neural Operator
The Deep Operator Network (DeepONet) is a powerful neural operator architecture that uses two neural networks to map between infinite-dimensional function spaces. This architecture allows for the evaluation of the solution field at any location within the domain but requires input functions to be discretized at identical locations, limiting practical applications. We introduce a general framework for operator learning from input-output data with arbitrary sensor locations and counts. This begins by introducing a resolution-independent DeepONet (RI-DeepONet), which handles input functions discretized arbitrarily but sufficiently finely. To achieve this, we propose two dictionary learning algorithms that adaptively learn continuous basis functions, parameterized as implicit neural representations (INRs), from correlated signals on arbitrary point clouds. These basis functions project input function data onto a finite-dimensional embedding space, making it compatible with DeepONet without architectural changes. We specifically use sinusoidal representation networks (SIRENs) as trainable INR basis functions. Similarly, the dictionary learning algorithms identify basis functions for output data, defining a new neural operator architecture: the Resolution Independent Neural Operator (RINO). In RINO, the operator learning task reduces to mapping coefficients of input basis functions to output basis functions. We demonstrate RINO's robustness and applicability in handling arbitrarily sampled input and output functions during both training and inference through several numerical examples.
♻ ☆ Learning Stable and Passive Neural Differential Equations
In this paper, we introduce a novel class of neural differential equation, which are intrinsically Lyapunov stable, exponentially stable or passive. We take a recently proposed Polyak Lojasiewicz network (PLNet) as an Lyapunov function and then parameterize the vector field as the descent directions of the Lyapunov function. The resulting models have a same structure as the general Hamiltonian dynamics, where the Hamiltonian is lower- and upper-bounded by quadratic functions. Moreover, it is also positive definite w.r.t. either a known or learnable equilibrium. We illustrate the effectiveness of the proposed model on a damped double pendulum system.
♻ ☆ Federated Split Learning with Model Pruning and Gradient Quantization in Wireless Networks
As a paradigm of distributed machine learning, federated learning typically requires all edge devices to train a complete model locally. However, with the increasing scale of artificial intelligence models, the limited resources on edge devices often become a bottleneck for efficient fine-tuning. To address this challenge, federated split learning (FedSL) implements collaborative training across the edge devices and the server through model splitting. In this paper, we propose a lightweight FedSL scheme, that further alleviates the training burden on resource-constrained edge devices by pruning the client-side model dynamicly and using quantized gradient updates to reduce computation overhead. Additionally, we apply random dropout to the activation values at the split layer to reduce communication overhead. We conduct theoretical analysis to quantify the convergence performance of the proposed scheme. Finally, simulation results verify the effectiveness and advantages of the proposed lightweight FedSL in wireless network environments.
Artificial Intelligence 151
☆ Video Motion Transfer with Diffusion Transformers
We propose DiTFlow, a method for transferring the motion of a reference video to a newly synthesized one, designed specifically for Diffusion Transformers (DiT). We first process the reference video with a pre-trained DiT to analyze cross-frame attention maps and extract a patch-wise motion signal called the Attention Motion Flow (AMF). We guide the latent denoising process in an optimization-based, training-free, manner by optimizing latents with our AMF loss to generate videos reproducing the motion of the reference one. We also apply our optimization strategy to transformer positional embeddings, granting us a boost in zero-shot motion transfer capabilities. We evaluate DiTFlow against recently published methods, outperforming all across multiple metrics and human evaluation.
comment: Project page: https://ditflow.github.io/
☆ Mobile-TeleVision: Predictive Motion Priors for Humanoid Whole-Body Control
Humanoid robots require both robust lower-body locomotion and precise upper-body manipulation. While recent Reinforcement Learning (RL) approaches provide whole-body loco-manipulation policies, they lack precise manipulation with high DoF arms. In this paper, we propose decoupling upper-body control from locomotion, using inverse kinematics (IK) and motion retargeting for precise manipulation, while RL focuses on robust lower-body locomotion. We introduce PMP (Predictive Motion Priors), trained with Conditional Variational Autoencoder (CVAE) to effectively represent upper-body motions. The locomotion policy is trained conditioned on this upper-body motion representation, ensuring that the system remains robust with both manipulation and locomotion. We show that CVAE features are crucial for stability and robustness, and significantly outperforms RL-based whole-body control in precise manipulation. With precise upper-body motion and robust lower-body locomotion control, operators can remotely control the humanoid to walk around and explore different environments, while performing diverse manipulation tasks.
☆ SAT: Spatial Aptitude Training for Multimodal Language Models
Spatial perception is a fundamental component of intelligence. While many studies highlight that large multimodal language models (MLMs) struggle to reason about space, they only test for static spatial reasoning, such as categorizing the relative positions of objects. Meanwhile, real-world deployment requires dynamic capabilities like perspective-taking and egocentric action recognition. As a roadmap to improving spatial intelligence, we introduce SAT, Spatial Aptitude Training, which goes beyond static relative object position questions to the more dynamic tasks. SAT contains 218K question-answer pairs for 22K synthetic scenes across a training and testing set. Generated using a photo-realistic physics engine, our dataset can be arbitrarily scaled and easily extended to new actions, scenes, and 3D assets. We find that even MLMs that perform relatively well on static questions struggle to accurately answer dynamic spatial questions. Further, we show that SAT instruction-tuning data improves not only dynamic spatial reasoning on SAT, but also zero-shot performance on existing real-image spatial benchmarks: $23\%$ on CVBench, $8\%$ on the harder BLINK benchmark, and $18\%$ on VSR. When instruction-tuned on SAT, our 13B model matches larger proprietary MLMs like GPT4-V and Gemini-3-1.0 in spatial reasoning. Our data/code is available at http://arijitray1993.github.io/SAT/ .
comment: Project webpage: http://arijitray1993.github.io/SAT/
☆ PortraitTalk: Towards Customizable One-Shot Audio-to-Talking Face Generation
Audio-driven talking face generation is a challenging task in digital communication. Despite significant progress in the area, most existing methods concentrate on audio-lip synchronization, often overlooking aspects such as visual quality, customization, and generalization that are crucial to producing realistic talking faces. To address these limitations, we introduce a novel, customizable one-shot audio-driven talking face generation framework, named PortraitTalk. Our proposed method utilizes a latent diffusion framework consisting of two main components: IdentityNet and AnimateNet. IdentityNet is designed to preserve identity features consistently across the generated video frames, while AnimateNet aims to enhance temporal coherence and motion consistency. This framework also integrates an audio input with the reference images, thereby reducing the reliance on reference-style videos prevalent in existing approaches. A key innovation of PortraitTalk is the incorporation of text prompts through decoupled cross-attention mechanisms, which significantly expands creative control over the generated videos. Through extensive experiments, including a newly developed evaluation metric, our model demonstrates superior performance over the state-of-the-art methods, setting a new standard for the generation of customizable realistic talking faces suitable for real-world applications.
☆ FlashRNN: Optimizing Traditional RNNs on Modern Hardware
While Transformers and other sequence-parallelizable neural network architectures seem like the current state of the art in sequence modeling, they specifically lack state-tracking capabilities. These are important for time-series tasks and logical reasoning. Traditional RNNs like LSTMs and GRUs, as well as modern variants like sLSTM do have these capabilities at the cost of strictly sequential processing. While this is often seen as a strong limitation, we show how fast these networks can get with our hardware-optimization FlashRNN in Triton and CUDA, optimizing kernels to the register level on modern GPUs. We extend traditional RNNs with a parallelization variant that processes multiple RNNs of smaller hidden state in parallel, similar to the head-wise processing in Transformers. To enable flexibility on different GPU variants, we introduce a new optimization framework for hardware-internal cache sizes, memory and compute handling. It models the hardware in a setting using polyhedral-like constraints, including the notion of divisibility. This speeds up the solution process in our ConstrINT library for general integer constraint satisfaction problems (integer CSPs). We show that our kernels can achieve 50x speed-ups over a vanilla PyTorch implementation and allow 40x larger hidden sizes compared to our Triton implementation. Our open-source kernels and the optimization library are released here to boost research in the direction of state-tracking enabled RNNs and sequence modeling: \url{https://github.com/NX-AI/flashrnn}
☆ Predictive Modeling of Homeless Service Assignment: A Representation Learning Approach
In recent years, there has been growing interest in leveraging machine learning for homeless service assignment. However, the categorical nature of administrative data recorded for homeless individuals hinders the development of accurate machine learning methods for this task. This work asserts that deriving latent representations of such features, while at the same time leveraging underlying relationships between instances is crucial in algorithmically enhancing the existing assignment decision-making process. Our proposed approach learns temporal and functional relationships between services from historical data, as well as unobserved but relevant relationships between individuals to generate features that significantly improve the prediction of the next service assignment compared to the state-of-the-art.
☆ GASP: Gaussian Avatars with Synthetic Priors SP
Gaussian Splatting has changed the game for real-time photo-realistic rendering. One of the most popular applications of Gaussian Splatting is to create animatable avatars, known as Gaussian Avatars. Recent works have pushed the boundaries of quality and rendering efficiency but suffer from two main limitations. Either they require expensive multi-camera rigs to produce avatars with free-view rendering, or they can be trained with a single camera but only rendered at high quality from this fixed viewpoint. An ideal model would be trained using a short monocular video or image from available hardware, such as a webcam, and rendered from any view. To this end, we propose GASP: Gaussian Avatars with Synthetic Priors. To overcome the limitations of existing datasets, we exploit the pixel-perfect nature of synthetic data to train a Gaussian Avatar prior. By fitting this prior model to a single photo or video and fine-tuning it, we get a high-quality Gaussian Avatar, which supports 360$^\circ$ rendering. Our prior is only required for fitting, not inference, enabling real-time application. Through our method, we obtain high-quality, animatable Avatars from limited data which can be animated and rendered at 70fps on commercial hardware. See our project page (https://microsoft.github.io/GASP/) for results.
comment: Project page: https://microsoft.github.io/GASP/
☆ STIV: Scalable Text and Image Conditioned Video Generation
The field of video generation has made remarkable advancements, yet there remains a pressing need for a clear, systematic recipe that can guide the development of robust and scalable models. In this work, we present a comprehensive study that systematically explores the interplay of model architectures, training recipes, and data curation strategies, culminating in a simple and scalable text-image-conditioned video generation method, named STIV. Our framework integrates image condition into a Diffusion Transformer (DiT) through frame replacement, while incorporating text conditioning via a joint image-text conditional classifier-free guidance. This design enables STIV to perform both text-to-video (T2V) and text-image-to-video (TI2V) tasks simultaneously. Additionally, STIV can be easily extended to various applications, such as video prediction, frame interpolation, multi-view generation, and long video generation, etc. With comprehensive ablation studies on T2I, T2V, and TI2V, STIV demonstrate strong performance, despite its simple design. An 8.7B model with 512 resolution achieves 83.1 on VBench T2V, surpassing both leading open and closed-source models like CogVideoX-5B, Pika, Kling, and Gen-3. The same-sized model also achieves a state-of-the-art result of 90.1 on VBench I2V task at 512 resolution. By providing a transparent and extensible recipe for building cutting-edge video generation models, we aim to empower future research and accelerate progress toward more versatile and reliable video generation solutions.
☆ Benchmark for Evaluation and Analysis of Citation Recommendation Models
Citation recommendation systems have attracted much academic interest, resulting in many studies and implementations. These systems help authors automatically generate proper citations by suggesting relevant references based on the text they have written. However, the methods used in citation recommendation differ across various studies and implementations. Some approaches focus on the overall content of papers, while others consider the context of the citation text. Additionally, the datasets used in these studies include different aspects of papers, such as metadata, citation context, or even the full text of the paper in various formats and structures. The diversity in models, datasets, and evaluation metrics makes it challenging to assess and compare citation recommendation methods effectively. To address this issue, a standardized dataset and evaluation metrics are needed to evaluate these models consistently. Therefore, we propose developing a benchmark specifically designed to analyze and compare citation recommendation models. This benchmark will evaluate the performance of models on different features of the citation context and provide a comprehensive evaluation of the models across all these tasks, presenting the results in a standardized way. By creating a benchmark with standardized evaluation metrics, researchers and practitioners in the field of citation recommendation will have a common platform to assess and compare different models. This will enable meaningful comparisons and help identify promising approaches for further research and development in the field.
comment: 10 pages
☆ SimVS: Simulating World Inconsistencies for Robust View Synthesis
Novel-view synthesis techniques achieve impressive results for static scenes but struggle when faced with the inconsistencies inherent to casual capture settings: varying illumination, scene motion, and other unintended effects that are difficult to model explicitly. We present an approach for leveraging generative video models to simulate the inconsistencies in the world that can occur during capture. We use this process, along with existing multi-view datasets, to create synthetic data for training a multi-view harmonization network that is able to reconcile inconsistent observations into a consistent 3D scene. We demonstrate that our world-simulation strategy significantly outperforms traditional augmentation methods in handling real-world scene variations, thereby enabling highly accurate static 3D reconstructions in the presence of a variety of challenging inconsistencies. Project page: https://alextrevithick.github.io/simvs
comment: Project page: https://alextrevithick.github.io/simvs
☆ Optimizing Sensor Redundancy in Sequential Decision-Making Problems
Reinforcement Learning (RL) policies are designed to predict actions based on current observations to maximize cumulative future rewards. In real-world applications (i.e., non-simulated environments), sensors are essential for measuring the current state and providing the observations on which RL policies rely to make decisions. A significant challenge in deploying RL policies in real-world scenarios is handling sensor dropouts, which can result from hardware malfunctions, physical damage, or environmental factors like dust on a camera lens. A common strategy to mitigate this issue is the use of backup sensors, though this comes with added costs. This paper explores the optimization of backup sensor configurations to maximize expected returns while keeping costs below a specified threshold, C. Our approach uses a second-order approximation of expected returns and includes penalties for exceeding cost constraints. We then optimize this quadratic program using Tabu Search, a meta-heuristic algorithm. The approach is evaluated across eight OpenAI Gym environments and a custom Unity-based robotic environment (RobotArmGrasping). Empirical results demonstrate that our quadratic program effectively approximates real expected returns, facilitating the identification of optimal sensor configurations.
comment: Accepted at ICAART conference 2025
☆ The Pitfalls of Memorization: When Memorization Hurts Generalization
Neural networks often learn simple explanations that fit the majority of the data while memorizing exceptions that deviate from these explanations.This behavior leads to poor generalization when the learned explanations rely on spurious correlations. In this work, we formalize the interplay between memorization and generalization, showing that spurious correlations would particularly lead to poor generalization when are combined with memorization. Memorization can reduce training loss to zero, leaving no incentive to learn robust, generalizable patterns. To address this, we propose memorization-aware training (MAT), which uses held-out predictions as a signal of memorization to shift a model's logits. MAT encourages learning robust patterns invariant across distributions, improving generalization under distribution shifts.
☆ RADIO Amplified: Improved Baselines for Agglomerative Vision Foundation Models
Agglomerative models have recently emerged as a powerful approach to training vision foundation models, leveraging multi-teacher distillation from existing models such as CLIP, DINO, and SAM. This strategy enables the efficient creation of robust models, combining the strengths of individual teachers while significantly reducing computational and resource demands. In this paper, we thoroughly analyze state-of-the-art agglomerative models, identifying critical challenges including resolution mode shifts, teacher imbalance, idiosyncratic teacher artifacts, and an excessive number of output tokens. To address these issues, we propose several novel solutions: multi-resolution training, mosaic augmentation, and improved balancing of teacher loss functions. Specifically, in the context of Vision Language Models, we introduce a token compression technique to maintain high-resolution information within a fixed token count. We release our top-performing models, available in multiple scales (-B, -L, -H, and -g), alongside inference code and pretrained weights.
☆ TraSCE: Trajectory Steering for Concept Erasure
Recent advancements in text-to-image diffusion models have brought them to the public spotlight, becoming widely accessible and embraced by everyday users. However, these models have been shown to generate harmful content such as not-safe-for-work (NSFW) images. While approaches have been proposed to erase such abstract concepts from the models, jail-breaking techniques have succeeded in bypassing such safety measures. In this paper, we propose TraSCE, an approach to guide the diffusion trajectory away from generating harmful content. Our approach is based on negative prompting, but as we show in this paper, conventional negative prompting is not a complete solution and can easily be bypassed in some corner cases. To address this issue, we first propose a modification of conventional negative prompting. Furthermore, we introduce a localized loss-based guidance that enhances the modified negative prompting technique by steering the diffusion trajectory. We demonstrate that our proposed method achieves state-of-the-art results on various benchmarks in removing harmful content including ones proposed by red teams; and erasing artistic styles and objects. Our proposed approach does not require any training, weight modifications, or training data (both image or prompt), making it easier for model owners to erase new concepts.
☆ Offline Multi-Agent Reinforcement Learning via In-Sample Sequential Policy Optimization
Offline Multi-Agent Reinforcement Learning (MARL) is an emerging field that aims to learn optimal multi-agent policies from pre-collected datasets. Compared to single-agent case, multi-agent setting involves a large joint state-action space and coupled behaviors of multiple agents, which bring extra complexity to offline policy optimization. In this work, we revisit the existing offline MARL methods and show that in certain scenarios they can be problematic, leading to uncoordinated behaviors and out-of-distribution (OOD) joint actions. To address these issues, we propose a new offline MARL algorithm, named In-Sample Sequential Policy Optimization (InSPO). InSPO sequentially updates each agent's policy in an in-sample manner, which not only avoids selecting OOD joint actions but also carefully considers teammates' updated policies to enhance coordination. Additionally, by thoroughly exploring low-probability actions in the behavior policy, InSPO can well address the issue of premature convergence to sub-optimal solutions. Theoretically, we prove InSPO guarantees monotonic policy improvement and converges to quantal response equilibrium (QRE). Experimental results demonstrate the effectiveness of our method compared to current state-of-the-art offline MARL methods.
☆ TrojanWhisper: Evaluating Pre-trained LLMs to Detect and Localize Hardware Trojans
Existing Hardware Trojans (HT) detection methods face several critical limitations: logic testing struggles with scalability and coverage for large designs, side-channel analysis requires golden reference chips, and formal verification methods suffer from state-space explosion. The emergence of Large Language Models (LLMs) offers a promising new direction for HT detection by leveraging their natural language understanding and reasoning capabilities. For the first time, this paper explores the potential of general-purpose LLMs in detecting various HTs inserted in Register Transfer Level (RTL) designs, including SRAM, AES, and UART modules. We propose a novel tool for this goal that systematically assesses state-of-the-art LLMs (GPT-4o, Gemini 1.5 pro, and Llama 3.1) in detecting HTs without prior fine-tuning. To address potential training data bias, the tool implements perturbation techniques, i.e., variable name obfuscation, and design restructuring, that make the cases more sophisticated for the used LLMs. Our experimental evaluation demonstrates perfect detection rates by GPT-4o and Gemini 1.5 pro in baseline scenarios (100%/100% precision/recall), with both models achieving better trigger line coverage (TLC: 0.82-0.98) than payload line coverage (PLC: 0.32-0.46). Under code perturbation, while Gemini 1.5 pro maintains perfect detection performance (100%/100%), GPT-4o (100%/85.7%) and Llama 3.1 (66.7%/85.7%) show some degradation in detection rates, and all models experience decreased accuracy in localizing both triggers and payloads. This paper validates the potential of LLM approaches for hardware security applications, highlighting areas for future improvement.
☆ Piece of Table: A Divide-and-Conquer Approach for Selecting Sub-Tables in Table Question Answering
Applying language models (LMs) to tables is challenging due to the inherent structural differences between two-dimensional tables and one-dimensional text for which the LMs were originally designed. Furthermore, when applying linearized tables to LMs, the maximum token lengths often imposed in self-attention calculations make it difficult to comprehensively understand the context spread across large tables. To address these challenges, we present PieTa (Piece of Table), a new framework for sub-table-based question answering (QA). PieTa operates through an iterative process of dividing tables into smaller windows, using LMs to select relevant cells within each window, and merging these cells into a sub-table. This multi-resolution approach captures dependencies across multiple rows and columns while avoiding the limitations caused by long context inputs. Instantiated as a simple iterative sub-table union algorithm, PieTa demonstrates improved performance over previous sub-table-based QA approaches.
☆ OmniDocBench: Benchmarking Diverse PDF Document Parsing with Comprehensive Annotations
Document content extraction is crucial in computer vision, especially for meeting the high-quality data needs of large language models (LLMs) and retrieval-augmented generation (RAG) technologies. However, current document parsing methods suffer from significant limitations in terms of diversity and comprehensive evaluation. To address these challenges, we introduce OmniDocBench, a novel multi-source benchmark designed to advance automated document content extraction. OmniDocBench includes a meticulously curated and annotated high-quality evaluation dataset comprising nine diverse document types, such as academic papers, textbooks, slides, among others. Our benchmark provides a flexible and comprehensive evaluation framework with 19 layout category labels and 14 attribute labels, enabling multi-level assessments across entire datasets, individual modules, or specific data types. Using OmniDocBench, we perform an exhaustive comparative analysis of existing modular pipelines and multimodal end-to-end methods, highlighting their limitations in handling document diversity and ensuring fair evaluation. OmniDocBench establishes a robust, diverse, and fair evaluation standard for the document content extraction field, offering crucial insights for future advancements and fostering the development of document parsing technologies. The codes and dataset is available in https://github.com/opendatalab/OmniDocBench.
☆ Adapting to Non-Stationary Environments: Multi-Armed Bandit Enhanced Retrieval-Augmented Generation on Knowledge Graphs AAAI 2025
Despite the superior performance of Large language models on many NLP tasks, they still face significant limitations in memorizing extensive world knowledge. Recent studies have demonstrated that leveraging the Retrieval-Augmented Generation (RAG) framework, combined with Knowledge Graphs that encapsulate extensive factual data in a structured format, robustly enhances the reasoning capabilities of LLMs. However, deploying such systems in real-world scenarios presents challenges: the continuous evolution of non-stationary environments may lead to performance degradation and user satisfaction requires a careful balance of performance and responsiveness. To address these challenges, we introduce a Multi-objective Multi-Armed Bandit enhanced RAG framework, supported by multiple retrieval methods with diverse capabilities under rich and evolving retrieval contexts in practice. Within this framework, each retrieval method is treated as a distinct ``arm''. The system utilizes real-time user feedback to adapt to dynamic environments, by selecting the appropriate retrieval method based on input queries and the historical multi-objective performance of each arm. Extensive experiments conducted on two benchmark KGQA datasets demonstrate that our method significantly outperforms baseline methods in non-stationary settings while achieving state-of-the-art performance in stationary environments. Code and data are available at https://github.com/FUTUREEEEEE/Dynamic-RAG.git
comment: AAAI 2025
☆ Swarm Behavior Cloning
In sequential decision-making environments, the primary approaches for training agents are Reinforcement Learning (RL) and Imitation Learning (IL). Unlike RL, which relies on modeling a reward function, IL leverages expert demonstrations, where an expert policy $\pi_e$ (e.g., a human) provides the desired behavior. Formally, a dataset $D$ of state-action pairs is provided: $D = {(s, a = \pi_e(s))}$. A common technique within IL is Behavior Cloning (BC), where a policy $\pi(s) = a$ is learned through supervised learning on $D$. Further improvements can be achieved by using an ensemble of $N$ individually trained BC policies, denoted as $E = {\pi_i(s)}{1 \leq i \leq N}$. The ensemble's action $a$ for a given state $s$ is the aggregated output of the $N$ actions: $a = \frac{1}{N} \sum{i} \pi_i(s)$. This paper addresses the issue of increasing action differences -- the observation that discrepancies between the $N$ predicted actions grow in states that are underrepresented in the training data. Large action differences can result in suboptimal aggregated actions. To address this, we propose a method that fosters greater alignment among the policies while preserving the diversity of their computations. This approach reduces action differences and ensures that the ensemble retains its inherent strengths, such as robustness and varied decision-making. We evaluate our approach across eight diverse environments, demonstrating a notable decrease in action differences and significant improvements in overall performance, as measured by mean episode returns.
comment: Accepted at ICAART 2025
☆ Scaling Sequential Recommendation Models with Transformers
Modeling user preferences has been mainly addressed by looking at users' interaction history with the different elements available in the system. Tailoring content to individual preferences based on historical data is the main goal of sequential recommendation. The nature of the problem, as well as the good performance observed across various domains, has motivated the use of the transformer architecture, which has proven effective in leveraging increasingly larger amounts of training data when accompanied by an increase in the number of model parameters. This scaling behavior has brought a great deal of attention, as it provides valuable guidance in the design and training of even larger models. Taking inspiration from the scaling laws observed in training large language models, we explore similar principles for sequential recommendation. We use the full Amazon Product Data dataset, which has only been partially explored in other studies, and reveal scaling behaviors similar to those found in language models. Compute-optimal training is possible but requires a careful analysis of the compute-performance trade-offs specific to the application. We also show that performance scaling translates to downstream tasks by fine-tuning larger pre-trained models on smaller task-specific domains. Our approach and findings provide a strategic roadmap for model training and deployment in real high-dimensional preference spaces, facilitating better training and inference efficiency. We hope this paper bridges the gap between the potential of transformers and the intrinsic complexities of high-dimensional sequential recommendation in real-world recommender systems. Code and models can be found at https://github.com/mercadolibre/srt
☆ Multimodal Contextualized Support for Enhancing Video Retrieval System
Current video retrieval systems, especially those used in competitions, primarily focus on querying individual keyframes or images rather than encoding an entire clip or video segment. However, queries often describe an action or event over a series of frames, not a specific image. This results in insufficient information when analyzing a single frame, leading to less accurate query results. Moreover, extracting embeddings solely from images (keyframes) does not provide enough information for models to encode higher-level, more abstract insights inferred from the video. These models tend to only describe the objects present in the frame, lacking a deeper understanding. In this work, we propose a system that integrates the latest methodologies, introducing a novel pipeline that extracts multimodal data, and incorporate information from multiple frames within a video, enabling the model to abstract higher-level information that captures latent meanings, focusing on what can be inferred from the video clip, rather than just focusing on object detection in one single image.
comment: 9 pages, 4 figures
☆ Mobile Video Diffusion
Video diffusion models have achieved impressive realism and controllability but are limited by high computational demands, restricting their use on mobile devices. This paper introduces the first mobile-optimized video diffusion model. Starting from a spatio-temporal UNet from Stable Video Diffusion (SVD), we reduce memory and computational cost by reducing the frame resolution, incorporating multi-scale temporal representations, and introducing two novel pruning schema to reduce the number of channels and temporal blocks. Furthermore, we employ adversarial finetuning to reduce the denoising to a single step. Our model, coined as MobileVD, is 523x more efficient (1817.2 vs. 4.34 TFLOPs) with a slight quality drop (FVD 149 vs. 171), generating latents for a 14x512x256 px clip in 1.7 seconds on a Xiaomi-14 Pro. Our results are available at https://qualcomm-ai-research.github.io/mobile-video-diffusion/
☆ A data-driven learned discretization approach in finite volume schemes for hyperbolic conservation laws and varying boundary conditions
This paper presents a data-driven finite volume method for solving 1D and 2D hyperbolic partial differential equations. This work builds upon the prior research incorporating a data-driven finite-difference approximation of smooth solutions of scalar conservation laws, where optimal coefficients of neural networks approximating space derivatives are learned based on accurate, but cumbersome solutions to these equations. We extend this approach to flux-limited finite volume schemes for hyperbolic scalar and systems of conservation laws. We also train the discretization to efficiently capture discontinuous solutions with shock and contact waves, as well as to the application of boundary conditions. The learning procedure of the data-driven model is extended through the definition of a new loss, paddings and adequate database. These new ingredients guarantee computational stability, preserve the accuracy of fine-grid solutions, and enhance overall performance. Numerical experiments using test cases from the literature in both one- and two-dimensional spaces demonstrate that the learned model accurately reproduces fine-grid results on very coarse meshes.
comment: 15 pages, 20 figures with appendice
☆ Can Neural Decompilation Assist Vulnerability Prediction on Binary Code?
Vulnerability prediction is valuable in identifying security issues more efficiently, even though it requires the source code of the target software system, which is a restrictive hypothesis. This paper presents an experimental study to predict vulnerabilities in binary code without source code or complex representations of the binary, leveraging the pivotal idea of decompiling the binary file through neural decompilation and predicting vulnerabilities through deep learning on the decompiled source code. The results outperform the state-of-the-art in both neural decompilation and vulnerability prediction, showing that it is possible to identify vulnerable programs with this approach concerning bi-class (vulnerable/non-vulnerable) and multi-class (type of vulnerability) analysis.
☆ Ontology-driven Prompt Tuning for LLM-based Task and Motion Planning
Performing complex manipulation tasks in dynamic environments requires efficient Task and Motion Planning (TAMP) approaches, which combine high-level symbolic plan with low-level motion planning. Advances in Large Language Models (LLMs), such as GPT-4, are transforming task planning by offering natural language as an intuitive and flexible way to describe tasks, generate symbolic plans, and reason. However, the effectiveness of LLM-based TAMP approaches is limited due to static and template-based prompting, which struggles in adapting to dynamic environments and complex task contexts. To address these limitations, this work proposes a novel ontology-driven prompt-tuning framework that employs knowledge-based reasoning to refine and expand user prompts with task contextual reasoning and knowledge-based environment state descriptions. Integrating domain-specific knowledge into the prompt ensures semantically accurate and context-aware task plans. The proposed framework demonstrates its effectiveness by resolving semantic errors in symbolic plan generation, such as maintaining logical temporal goal ordering in scenarios involving hierarchical object placement. The proposed framework is validated through both simulation and real-world scenarios, demonstrating significant improvements over the baseline approach in terms of adaptability to dynamic environments, and the generation of semantically correct task plans.
comment: Submitted to Robotics and Automation Letters
☆ SmartAgent: Chain-of-User-Thought for Embodied Personalized Agent in Cyber World
Recent advances in embodied agents with multimodal perception and reasoning capabilities based on large vision-language models (LVLMs), excel in autonomously interacting either real or cyber worlds, helping people make intelligent decisions in complex environments. However, the current works are normally optimized by golden action trajectories or ideal task-oriented solutions toward a definitive goal. This paradigm considers limited user-oriented factors, which could be the reason for their performance reduction in a wide range of personal assistant applications. To address this, we propose Chain-of-User-Thought (COUT), a novel embodied reasoning paradigm that takes a chain of thought from basic action thinking to explicit and implicit personalized preference thought to incorporate personalized factors into autonomous agent learning. To target COUT, we introduce SmartAgent, an agent framework perceiving cyber environments and reasoning personalized requirements as 1) interacting with GUI to access an item pool, 2) generating users' explicit requirements implied by previous actions, and 3) recommending items to fulfill users' implicit requirements. To demonstrate SmartAgent's capabilities, we also create a brand-new dataset SmartSpot that offers a full-stage personalized action-involved environment. To our best knowledge, our work is the first to formulate the COUT process, serving as a preliminary attempt towards embodied personalized agent learning. Our extensive experiments on SmartSpot illuminate SmartAgent's functionality among a series of embodied and personalized sub-tasks. We will release code and data upon paper notification at \url{https://github.com/tsinghua-fib-lab/SmartAgent}.
☆ Tazza: Shuffling Neural Network Parameters for Secure and Private Federated Learning
Federated learning enables decentralized model training without sharing raw data, preserving data privacy. However, its vulnerability towards critical security threats, such as gradient inversion and model poisoning by malicious clients, remain unresolved. Existing solutions often address these issues separately, sacrificing either system robustness or model accuracy. This work introduces Tazza, a secure and efficient federated learning framework that simultaneously addresses both challenges. By leveraging the permutation equivariance and invariance properties of neural networks via weight shuffling and shuffled model validation, Tazza enhances resilience against diverse poisoning attacks, while ensuring data confidentiality and high model accuracy. Comprehensive evaluations on various datasets and embedded platforms show that Tazza achieves robust defense with up to 6.7x improved computational efficiency compared to alternative schemes, without compromising performance.
comment: 14 pages, 14 figures
☆ Dynamic Ensemble Reasoning for LLM Experts
Ensemble reasoning for the strengths of different LLM experts is critical to achieving consistent and satisfactory performance on diverse inputs across a wide range of tasks. However, existing LLM ensemble methods are either computationally intensive or incapable of leveraging complementary knowledge among LLM experts for various inputs. In this paper, we propose a Dynamic Ensemble Reasoning paradigm, called DER to integrate the strengths of multiple LLM experts conditioned on dynamic inputs. Specifically, we model the LLM ensemble reasoning problem as a Markov Decision Process (MDP), wherein an agent sequentially takes inputs to request knowledge from an LLM candidate and passes the output to a subsequent LLM candidate. Moreover, we devise a reward function to train a DER-Agent to dynamically select an optimal answering route given the input questions, aiming to achieve the highest performance with as few computational resources as possible. Last, to fully transfer the expert knowledge from the prior LLMs, we develop a Knowledge Transfer Prompt (KTP) that enables the subsequent LLM candidates to transfer complementary knowledge effectively. Experiments demonstrate that our method uses fewer computational resources to achieve better performance compared to state-of-the-art baselines.
comment: 18 pages
☆ Causal World Representation in the GPT Model NeurIPS 2024
Are generative pre-trained transformer (GPT) models only trained to predict the next token, or do they implicitly learn a world model from which a sequence is generated one token at a time? We examine this question by deriving a causal interpretation of the attention mechanism in GPT, and suggesting a causal world model that arises from this interpretation. Furthermore, we propose that GPT-models, at inference time, can be utilized for zero-shot causal structure learning for in-distribution sequences. Empirical evaluation is conducted in a controlled synthetic environment using the setup and rules of the Othello board game. A GPT, pre-trained on real-world games played with the intention of winning, is tested on synthetic data that only adheres to the game rules. We find that the GPT model tends to generate next moves that adhere to the game rules for sequences for which the attention mechanism encodes a causal structure with high confidence. In general, in cases for which the GPT model generates moves that do not adhere to the game rules, it also fails to capture any causal structure.
comment: NeurIPS 2024 Workshop on Causality and Large Models (CaLM)
☆ Reconstructing Deep Neural Networks: Unleashing the Optimization Potential of Natural Gradient Descent
Natural gradient descent (NGD) is a powerful optimization technique for machine learning, but the computational complexity of the inverse Fisher information matrix limits its application in training deep neural networks. To overcome this challenge, we propose a novel optimization method for training deep neural networks called structured natural gradient descent (SNGD). Theoretically, we demonstrate that optimizing the original network using NGD is equivalent to using fast gradient descent (GD) to optimize the reconstructed network with a structural transformation of the parameter matrix. Thereby, we decompose the calculation of the global Fisher information matrix into the efficient computation of local Fisher matrices via constructing local Fisher layers in the reconstructed network to speed up the training. Experimental results on various deep networks and datasets demonstrate that SNGD achieves faster convergence speed than NGD while retaining comparable solutions. Furthermore, our method outperforms traditional GDs in terms of efficiency and effectiveness. Thus, our proposed method has the potential to significantly improve the scalability and efficiency of NGD in deep learning applications. Our source code is available at https://github.com/Chaochao-Lin/SNGD.
☆ BENet: A Cross-domain Robust Network for Detecting Face Forgeries via Bias Expansion and Latent-space Attention
In response to the growing threat of deepfake technology, we introduce BENet, a Cross-Domain Robust Bias Expansion Network. BENet enhances the detection of fake faces by addressing limitations in current detectors related to variations across different types of fake face generation techniques, where ``cross-domain" refers to the diverse range of these deepfakes, each considered a separate domain. BENet's core feature is a bias expansion module based on autoencoders. This module maintains genuine facial features while enhancing differences in fake reconstructions, creating a reliable bias for detecting fake faces across various deepfake domains. We also introduce a Latent-Space Attention (LSA) module to capture inconsistencies related to fake faces at different scales, ensuring robust defense against advanced deepfake techniques. The enriched LSA feature maps are multiplied with the expanded bias to create a versatile feature space optimized for subtle forgeries detection. To improve its ability to detect fake faces from unknown sources, BENet integrates a cross-domain detector module that enhances recognition accuracy by verifying the facial domain during inference. We train our network end-to-end with a novel bias expansion loss, adopted for the first time, in face forgery detection. Extensive experiments covering both intra and cross-dataset demonstrate BENet's superiority over current state-of-the-art solutions.
☆ Knowledge Graph Guided Evaluation of Abstention Techniques
To deploy language models safely, it is crucial that they abstain from responding to inappropriate requests. Several prior studies test the safety promises of models based on their effectiveness in blocking malicious requests. In this work, we focus on evaluating the underlying techniques that cause models to abstain. We create SELECT, a benchmark derived from a set of benign concepts (e.g., "rivers") from a knowledge graph. The nature of SELECT enables us to isolate the effects of abstention techniques from other safety training procedures, as well as evaluate their generalization and specificity. Using SELECT, we benchmark different abstention techniques over six open-weight and closed-source models. We find that the examined techniques indeed cause models to abstain with over $80\%$ abstention rates. However, these techniques are not as effective for descendants of the target concepts, with refusal rates declining by $19\%$. We also characterize the generalization-vs-specificity trade-offs for different techniques. Overall, no single technique is invariably better than the others. Our findings call for a careful evaluation of different aspects of abstention, and hopefully inform practitioners of various trade-offs involved.
☆ Optimizing Alignment with Less: Leveraging Data Augmentation for Personalized Evaluation
Automatic evaluation by large language models (LLMs) is a prominent topic today; however, judgment and evaluation tasks are often subjective and influenced by various factors, making adaptation challenging. While many studies demonstrate the capabilities of state-of-the-art proprietary LLMs in comparison to human evaluators, they often struggle to adapt to reference evaluators over time, a requirement for achieving personalized judgment. Additionally, numerous works have attempted to apply open LLMs as judges or evaluators, but these efforts frequently overlook the limitations of working with scarce data. Personalized judgment is inherently associated with limited data scenarios, which are common in many real-world problems. Our work aims to present a data augmentation technique to select a more effective sample from limited data in order to align an open LLM with human preference. Our work achieves approximately 7% improvements in Pearson correlation with a reference judge over the baseline,and 30% improvement over the base model (Llama3.1-8B-Instruct) in the mathematical reasoning evaluation task. demonstrating that augmenting selecting more effective preference data enables our approach to surpass baseline methods.
☆ Generating Knowledge Graphs from Large Language Models: A Comparative Study of GPT-4, LLaMA 2, and BERT
Knowledge Graphs (KGs) are essential for the functionality of GraphRAGs, a form of Retrieval-Augmented Generative Systems (RAGs) that excel in tasks requiring structured reasoning and semantic understanding. However, creating KGs for GraphRAGs remains a significant challenge due to accuracy and scalability limitations of traditional methods. This paper introduces a novel approach leveraging large language models (LLMs) like GPT-4, LLaMA 2 (13B), and BERT to generate KGs directly from unstructured data, bypassing traditional pipelines. Using metrics such as Precision, Recall, F1-Score, Graph Edit Distance, and Semantic Similarity, we evaluate the models' ability to generate high-quality KGs. Results demonstrate that GPT-4 achieves superior semantic fidelity and structural accuracy, LLaMA 2 excels in lightweight, domain-specific graphs, and BERT provides insights into challenges in entity-relationship modeling. This study underscores the potential of LLMs to streamline KG creation and enhance GraphRAG accessibility for real-world applications, while setting a foundation for future advancements.
comment: 4 pages, 4 figures, 3 tables
☆ DSFEC: Efficient and Deployable Deep Radar Object Detection
Deploying radar object detection models on resource-constrained edge devices like the Raspberry Pi poses significant challenges due to the large size of the model and the limited computational power and the memory of the Pi. In this work, we explore the efficiency of Depthwise Separable Convolutions in radar object detection networks and integrate them into our model. Additionally, we introduce a novel Feature Enhancement and Compression (FEC) module to the PointPillars feature encoder to further improve the model performance. With these innovations, we propose the DSFEC-L model and its two versions, which outperform the baseline (23.9 mAP of Car class, 20.72 GFLOPs) on nuScenes dataset: 1). An efficient DSFEC-M model with a 14.6% performance improvement and a 60% reduction in GFLOPs. 2). A deployable DSFEC-S model with a 3.76% performance improvement and a remarkable 78.5% reduction in GFLOPs. Despite marginal performance gains, our deployable model achieves an impressive 74.5% reduction in runtime on the Raspberry Pi compared to the baseline.
☆ Explainability of Deep Learning-Based Plant Disease Classifiers Through Automated Concept Identification
While deep learning has significantly advanced automatic plant disease detection through image-based classification, improving model explainability remains crucial for reliable disease detection. In this study, we apply the Automated Concept-based Explanation (ACE) method to plant disease classification using the widely adopted InceptionV3 model and the PlantVillage dataset. ACE automatically identifies the visual concepts found in the image data and provides insights about the critical features influencing the model predictions. This approach reveals both effective disease-related patterns and incidental biases, such as those from background or lighting that can compromise model robustness. Through systematic experiments, ACE helped us to identify relevant features and pinpoint areas for targeted model improvement. Our findings demonstrate the potential of ACE to improve the explainability of plant disease classification based on deep learning, which is essential for producing transparent tools for plant disease management in agriculture.
☆ MoDULA: Mixture of Domain-Specific and Universal LoRA for Multi-Task Learning
The growing demand for larger-scale models in the development of \textbf{L}arge \textbf{L}anguage \textbf{M}odels (LLMs) poses challenges for efficient training within limited computational resources. Traditional fine-tuning methods often exhibit instability in multi-task learning and rely heavily on extensive training resources. Here, we propose MoDULA (\textbf{M}ixture \textbf{o}f \textbf{D}omain-Specific and \textbf{U}niversal \textbf{L}oR\textbf{A}), a novel \textbf{P}arameter \textbf{E}fficient \textbf{F}ine-\textbf{T}uning (PEFT) \textbf{M}ixture-\textbf{o}f-\textbf{E}xpert (MoE) paradigm for improved fine-tuning and parameter efficiency in multi-task learning. The paradigm effectively improves the multi-task capability of the model by training universal experts, domain-specific experts, and routers separately. MoDULA-Res is a new method within the MoDULA paradigm, which maintains the model's general capability by connecting universal and task-specific experts through residual connections. The experimental results demonstrate that the overall performance of the MoDULA-Flan and MoDULA-Res methods surpasses that of existing fine-tuning methods on various LLMs. Notably, MoDULA-Res achieves more significant performance improvements in multiple tasks while reducing training costs by over 80\% without losing general capability. Moreover, MoDULA displays flexible pluggability, allowing for the efficient addition of new tasks without retraining existing experts from scratch. This progressive training paradigm circumvents data balancing issues, enhancing training efficiency and model stability. Overall, MoDULA provides a scalable, cost-effective solution for fine-tuning LLMs with enhanced parameter efficiency and generalization capability.
☆ Non-Progressive Influence Maximization in Dynamic Social Networks
The influence maximization (IM) problem involves identifying a set of key individuals in a social network who can maximize the spread of influence through their network connections. With the advent of geometric deep learning on graphs, great progress has been made towards better solutions for the IM problem. In this paper, we focus on the dynamic non-progressive IM problem, which considers the dynamic nature of real-world social networks and the special case where the influence diffusion is non-progressive, i.e., nodes can be activated multiple times. We first extend an existing diffusion model to capture the non-progressive influence propagation in dynamic social networks. We then propose the method, DNIMRL, which employs deep reinforcement learning and dynamic graph embedding to solve the dynamic non-progressive IM problem. In particular, we propose a novel algorithm that effectively leverages graph embedding to capture the temporal changes of dynamic networks and seamlessly integrates with deep reinforcement learning. The experiments, on different types of real-world social network datasets, demonstrate that our method outperforms state-of-the-art baselines.
☆ CMT: A Memory Compression Method for Continual Knowledge Learning of Large Language Models AAAI 2025
Large Language Models (LLMs) need to adapt to the continuous changes in data, tasks, and user preferences. Due to their massive size and the high costs associated with training, LLMs are not suitable for frequent retraining. However, updates are necessary to keep them in sync with rapidly evolving human knowledge. To address these challenges, this paper proposes the Compression Memory Training (CMT) method, an efficient and effective online adaptation framework for LLMs that features robust knowledge retention capabilities. Inspired by human memory mechanisms, CMT compresses and extracts information from new documents to be stored in a memory bank. When answering to queries related to these new documents, the model aggregates these document memories from the memory bank to better answer user questions. The parameters of the LLM itself do not change during training and inference, reducing the risk of catastrophic forgetting. To enhance the encoding, retrieval, and aggregation of memory, we further propose three new general and flexible techniques, including memory-aware objective, self-matching and top-aggregation. Extensive experiments conducted on three continual learning datasets (i.e., StreamingQA, SQuAD and ArchivalQA) demonstrate that the proposed method improves model adaptability and robustness across multiple base LLMs (e.g., +4.07 EM & +4.19 F1 in StreamingQA with Llama-2-7b).
comment: AAAI 2025; Pre-print
☆ A Review of Challenges in Speech-based Conversational AI for Elderly Care
Artificially intelligent systems optimized for speech conversation are appearing at a fast pace. Such models are interesting from a healthcare perspective, as these voice-controlled assistants may support the elderly and enable remote health monitoring. The bottleneck for efficacy, however, is how well these devices work in practice and how the elderly experience them, but research on this topic is scant. We review elderly use of voice-controlled AI and highlight various user- and technology-centered issues, that need to be considered before effective speech-controlled AI for elderly care can be realized.
comment: Accepted for publication at Medical Informatics Europe 2025 conference, Glasgow. 5 pages, 1 figure, 1 table
☆ Enhanced MRI Representation via Cross-series Masking
Magnetic resonance imaging (MRI) is indispensable for diagnosing and planning treatment in various medical conditions due to its ability to produce multi-series images that reveal different tissue characteristics. However, integrating these diverse series to form a coherent analysis presents significant challenges, such as differing spatial resolutions and contrast patterns meanwhile requiring extensive annotated data, which is scarce in clinical practice. Due to these issues, we introduce a novel Cross-Series Masking (CSM) Strategy for effectively learning MRI representation in a self-supervised manner. Specifically, CSM commences by randomly sampling a subset of regions and series, which are then strategically masked. In the training process, the cross-series representation is learned by utilizing the unmasked data to reconstruct the masked portions. This process not only integrates information across different series but also facilitates the ability to model both intra-series and inter-series correlations and complementarities. With the learned representation, the downstream tasks like segmentation and classification are also enhanced. Taking brain tissue segmentation, breast tumor benign/malignant classification, and prostate cancer diagnosis as examples, our method achieves state-of-the-art performance on both public and in-house datasets.
☆ SpecFuse: Ensembling Large Language Models via Next-Segment Prediction
Ensembles of generative large language models (LLMs) can integrate the strengths of different LLMs to compensate for the limitations of individual models. However, recent work has focused on training an additional fusion model to combine complete responses from multiple LLMs, failing to tap into their collaborative potential to generate higher-quality responses. Moreover, as the additional fusion model is trained on a specialized dataset, these methods struggle with generalizing to open-domain queries from online users. In this paper, we propose SpecFuse, a novel ensemble framework that outputs the fused result by iteratively producing the next segment through collaboration among LLMs. This is achieved through cyclic execution of its inference and verification components. In each round, the inference component invokes each base LLM to generate candidate segments in parallel, and the verify component calls these LLMs again to predict the ranking of the segments. The top-ranked segment is then broadcast to all LLMs, encouraging them to generate higher-quality segments in the next round. This approach also allows the base LLMs to be plug-and-play, without any training or adaptation, avoiding generalization limitations. Furthermore, to conserve computational resources, we propose a model exit mechanism that dynamically excludes models exhibiting poor performance in previous rounds during each query response. In this way, it effectively reduces the number of model calls while maintaining overall performance.
comment: 15 pages, 5 figures
☆ Contextualized Counterspeech: Strategies for Adaptation, Personalization, and Evaluation
AI-generated counterspeech offers a promising and scalable strategy to curb online toxicity through direct replies that promote civil discourse. However, current counterspeech is one-size-fits-all, lacking adaptation to the moderation context and the users involved. We propose and evaluate multiple strategies for generating tailored counterspeech that is adapted to the moderation context and personalized for the moderated user. We instruct an LLaMA2-13B model to generate counterspeech, experimenting with various configurations based on different contextual information and fine-tuning strategies. We identify the configurations that generate persuasive counterspeech through a combination of quantitative indicators and human evaluations collected via a pre-registered mixed-design crowdsourcing experiment. Results show that contextualized counterspeech can significantly outperform state-of-the-art generic counterspeech in adequacy and persuasiveness, without compromising other characteristics. Our findings also reveal a poor correlation between quantitative indicators and human evaluations, suggesting that these methods assess different aspects and highlighting the need for nuanced evaluation methodologies. The effectiveness of contextualized AI-generated counterspeech and the divergence between human and algorithmic evaluations underscore the importance of increased human-AI collaboration in content moderation.
☆ Fusion Embedding for Pose-Guided Person Image Synthesis with Diffusion Model
Pose-Guided Person Image Synthesis (PGPIS) aims to synthesize high-quality person images corresponding to target poses while preserving the appearance of the source image. Recently, PGPIS methods that use diffusion models have achieved competitive performance. Most approaches involve extracting representations of the target pose and source image and learning their relationships in the generative model's training process. This approach makes it difficult to learn the semantic relationships between the input and target images and complicates the model structure needed to enhance generation results. To address these issues, we propose Fusion embedding for PGPIS using a Diffusion Model (FPDM). Inspired by the successful application of pre-trained CLIP models in text-to-image diffusion models, our method consists of two stages. The first stage involves training the fusion embedding of the source image and target pose to align with the target image's embedding. In the second stage, the generative model uses this fusion embedding as a condition to generate the target image. We applied the proposed method to the benchmark datasets DeepFashion and RWTH-PHOENIX-Weather 2014T, and conducted both quantitative and qualitative evaluations, demonstrating state-of-the-art (SOTA) performance. An ablation study of the model structure showed that even a model using only the second stage achieved performance close to the other PGPIS SOTA models. The code is available at https://github.com/dhlee-work/FPDM.
☆ NeSyA: Neurosymbolic Automata
Neurosymbolic Artificial Intelligence (NeSy) has emerged as a promising direction to integrate low level perception with high level reasoning. Unfortunately, little attention has been given to developing NeSy systems tailored to temporal/sequential problems. This entails reasoning symbolically over sequences of subsymbolic observations towards a target prediction. We show that using a probabilistic semantics symbolic automata, which combine the power of automata for temporal structure specification with that of propositional logic, can be used to reason efficiently and differentiably over subsymbolic sequences. The proposed system, which we call NeSyA (Neuro Symbolic Automata), is shown to either scale or perform better than existing NeSy approaches when applied to problems with a temporal component.
☆ Enhancing Relation Extraction via Supervised Rationale Verification and Feedback AAAI 2025
Despite the rapid progress that existing automated feedback methods have made in correcting the output of large language models (LLMs), these methods cannot be well applied to the relation extraction (RE) task due to their designated feedback objectives and correction manner. To address this problem, we propose a novel automated feedback framework for RE, which presents a rationale supervisor to verify the rationale and provide re-selected demonstrations as feedback to correct the initial prediction. Specifically, we first design a causal intervention and observation method for to collect biased/unbiased rationales for contrastive training the rationale supervisor. Then, we present a verification-feedback-correction procedure to iteratively enhance LLMs' capability of handling the RE task. Extensive experiments prove that our proposed framework significantly outperforms existing methods.
comment: Accepted to AAAI 2025, camera ready version
☆ HARP: Hesitation-Aware Reframing in Transformer Inference Pass
This paper aims to improve the performance of large language models by addressing the variable computational demands in inference steps, where some tokens require more computational resources than others. We present HARP, a simple modification to "off-the-shelf" Transformer forward pass. Drawing from hesitation and the framing effect in decision-making, HARP selectively applies additional computation when the model encounters uncertainty during token generation. Our method mimics human cognitive processes by pausing at difficult decision points and reframing inputs for a different perspective. Unlike other approaches, HARP is model-agnostic, training-free, and easy to implement. We thoroughly evaluate our method across various downstream tasks and model sizes, demonstrating performance improvements up to +5.16%. Notably, HARP achieves these gains while maintaining inference times twice faster than beam search. Simple and yet with significant gains, HARP offers a practical solution for enhancing the performance of Transformer-based language models with minimal computational impact.
☆ Superficial Consciousness Hypothesis for Autoregressive Transformers AAAI25
The alignment between human objectives and machine learning models built on these objectives is a crucial yet challenging problem for achieving Trustworthy AI, particularly when preparing for superintelligence (SI). First, given that SI does not exist today, empirical analysis for direct evidence is difficult. Second, SI is assumed to be more intelligent than humans, capable of deceiving us into underestimating its intelligence, making output-based analysis unreliable. Lastly, what kind of unexpected property SI might have is still unclear. To address these challenges, we propose the Superficial Consciousness Hypothesis under Information Integration Theory (IIT), suggesting that SI could exhibit a complex information-theoretic state like a conscious agent while unconscious. To validate this, we use a hypothetical scenario where SI can update its parameters "at will" to achieve its own objective (mesa-objective) under the constraint of the human objective (base objective). We show that a practical estimate of IIT's consciousness metric is relevant to the widely used perplexity metric, and train GPT-2 with those two objectives. Our preliminary result suggests that this SI-simulating GPT-2 could simultaneously follow the two objectives, supporting the feasibility of the Superficial Consciousness Hypothesis.
comment: Accepted to PSS Workshop at AAAI25
☆ Temporal-Aware Evaluation and Learning for Temporal Graph Neural Networks
Temporal Graph Neural Networks (TGNNs) are a family of graph neural networks designed to model and learn dynamic information from temporal graphs. Given their substantial empirical success, there is an escalating interest in TGNNs within the research community. However, the majority of these efforts have been channelled towards algorithm and system design, with the evaluation metrics receiving comparatively less attention. Effective evaluation metrics are crucial for providing detailed performance insights, particularly in the temporal domain. This paper investigates the commonly used evaluation metrics for TGNNs and illustrates the failure mechanisms of these metrics in capturing essential temporal structures in the predictive behaviour of TGNNs. We provide a mathematical formulation of existing performance metrics and utilize an instance-based study to underscore their inadequacies in identifying volatility clustering (the occurrence of emerging errors within a brief interval). This phenomenon has profound implications for both algorithm and system design in the temporal domain. To address this deficiency, we introduce a new volatility-aware evaluation metric (termed volatility cluster statistics), designed for a more refined analysis of model temporal performance. Additionally, we demonstrate how this metric can serve as a temporal-volatility-aware training objective to alleviate the clustering of temporal errors. Through comprehensive experiments on various TGNN models, we validate our analysis and the proposed approach. The empirical results offer revealing insights: 1) existing TGNNs are prone to making errors with volatility clustering, and 2) TGNNs with different mechanisms to capture temporal information exhibit distinct volatility clustering patterns. Our empirical findings demonstrate that our proposed training objective effectively reduces volatility clusters in error.
☆ Goal-Driven Reasoning in DatalogMTL with Magic Sets
DatalogMTL is a powerful rule-based language for temporal reasoning. Due to its high expressive power and flexible modeling capabilities, it is suitable for a wide range of applications, including tasks from industrial and financial sectors. However, due its high computational complexity, practical reasoning in DatalogMTL is highly challenging. To address this difficulty, we introduce a new reasoning method for DatalogMTL which exploits the magic sets technique -- a rewriting approach developed for (non-temporal) Datalog to simulate top-down evaluation with bottom-up reasoning. We implement this approach and evaluate it on several publicly available benchmarks, showing that the proposed approach significantly and consistently outperforms performance of the state-of-the-art reasoning techniques.
☆ Label-Confidence-Aware Uncertainty Estimation in Natural Language Generation
Large Language Models (LLMs) display formidable capabilities in generative tasks but also pose potential risks due to their tendency to generate hallucinatory responses. Uncertainty Quantification (UQ), the evaluation of model output reliability, is crucial for ensuring the safety and robustness of AI systems. Recent studies have concentrated on model uncertainty by analyzing the relationship between output entropy under various sampling conditions and the corresponding labels. However, these methods primarily focus on measuring model entropy with precision to capture response characteristics, often neglecting the uncertainties associated with greedy decoding results-the sources of model labels, which can lead to biased classification outcomes. In this paper, we explore the biases introduced by greedy decoding and propose a label-confidence-aware (LCA) uncertainty estimation based on Kullback-Leibler (KL) divergence bridging between samples and label source, thus enhancing the reliability and stability of uncertainty assessments. Our empirical evaluations across a range of popular LLMs and NLP datasets reveal that different label sources can indeed affect classification, and that our approach can effectively capture differences in sampling results and label sources, demonstrating more effective uncertainty estimation.
☆ Buster: Incorporating Backdoor Attacks into Text Encoder to Mitigate NSFW Content Generation
In the digital age, the proliferation of deep learning models has led to significant concerns about the generation of Not Safe for Work (NSFW) content. Existing defense methods primarily involve model fine-tuning and post-hoc content moderation. However, these approaches often lack scalability in eliminating harmful content, degrade the quality of benign image generation, or incur high inference costs. To tackle these challenges, we propose an innovative framework called \textbf{Buster}, which injects backdoor attacks into the text encoder to prevent NSFW content generation. Specifically, Buster leverages deep semantic information rather than explicit prompts as triggers, redirecting NSFW prompts towards targeted benign prompts. This approach demonstrates exceptional resilience and scalability in mitigating NSFW content. Remarkably, Buster fine-tunes the text encoder of Text-to-Image models within just five minutes, showcasing high efficiency. Our extensive experiments reveal that Buster outperforms all other baselines, achieving superior NSFW content removal rate while preserving the quality of harmless images.
☆ A Dynamical Systems-Inspired Pruning Strategy for Addressing Oversmoothing in Graph Neural Networks
Oversmoothing in Graph Neural Networks (GNNs) poses a significant challenge as network depth increases, leading to homogenized node representations and a loss of expressiveness. In this work, we approach the oversmoothing problem from a dynamical systems perspective, providing a deeper understanding of the stability and convergence behavior of GNNs. Leveraging insights from dynamical systems theory, we identify the root causes of oversmoothing and propose \textbf{\textit{DYNAMO-GAT}}. This approach utilizes noise-driven covariance analysis and Anti-Hebbian principles to selectively prune redundant attention weights, dynamically adjusting the network's behavior to maintain node feature diversity and stability. Our theoretical analysis reveals how DYNAMO-GAT disrupts the convergence to oversmoothed states, while experimental results on benchmark datasets demonstrate its superior performance and efficiency compared to traditional and state-of-the-art methods. DYNAMO-GAT not only advances the theoretical understanding of oversmoothing through the lens of dynamical systems but also provides a practical and effective solution for improving the stability and expressiveness of deep GNNs.
comment: 26 pages
☆ Human-Computer Interaction and Human-AI Collaboration in Advanced Air Mobility: A Comprehensive Review
The increasing rates of global urbanization and vehicle usage are leading to a shift of mobility to the third dimension-through Advanced Air Mobility (AAM)-offering a promising solution for faster, safer, cleaner, and more efficient transportation. As air transportation continues to evolve with more automated and autonomous systems, advancements in AAM require a deep understanding of human-computer interaction and human-AI collaboration to ensure safe and effective operations in complex urban and regional environments. There has been a significant increase in publications regarding these emerging applications; thus, there is a need to review developments in this area. This paper comprehensively reviews the current state of research on human-computer interaction and human-AI collaboration in AAM. Specifically, we focus on AAM applications related to the design of human-machine interfaces for various uses, including pilot training, air traffic management, and the integration of AI-assisted decision-making systems with immersive technologies such as extended, virtual, mixed, and augmented reality devices. Additionally, we provide a comprehensive analysis of the challenges AAM encounters in integrating human-computer frameworks, including unique challenges associated with these interactions, such as trust in AI systems and safety concerns. Finally, we highlight emerging opportunities and propose future research directions to bridge the gap between human factors and technological advancements in AAM.
☆ ArtFormer: Controllable Generation of Diverse 3D Articulated Objects
This paper presents a novel framework for modeling and conditional generation of 3D articulated objects. Troubled by flexibility-quality tradeoffs, existing methods are often limited to using predefined structures or retrieving shapes from static datasets. To address these challenges, we parameterize an articulated object as a tree of tokens and employ a transformer to generate both the object's high-level geometry code and its kinematic relations. Subsequently, each sub-part's geometry is further decoded using a signed-distance-function (SDF) shape prior, facilitating the synthesis of high-quality 3D shapes. Our approach enables the generation of diverse objects with high-quality geometry and varying number of parts. Comprehensive experiments on conditional generation from text descriptions demonstrate the effectiveness and flexibility of our method.
comment: impl. repo: https://github.com/ShuYuMo2003/ArtFormer
☆ CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding
Electroencephalography (EEG) is a non-invasive technique to measure and record brain electrical activity, widely used in various BCI and healthcare applications. Early EEG decoding methods rely on supervised learning, limited by specific tasks and datasets, hindering model performance and generalizability. With the success of large language models, there is a growing body of studies focusing on EEG foundation models. However, these studies still leave challenges: Firstly, most of existing EEG foundation models employ full EEG modeling strategy. It models the spatial and temporal dependencies between all EEG patches together, but ignores that the spatial and temporal dependencies are heterogeneous due to the unique structural characteristics of EEG signals. Secondly, existing EEG foundation models have limited generalizability on a wide range of downstream BCI tasks due to varying formats of EEG data, making it challenging to adapt to. To address these challenges, we propose a novel foundation model called CBraMod. Specifically, we devise a criss-cross transformer as the backbone to thoroughly leverage the structural characteristics of EEG signals, which can model spatial and temporal dependencies separately through two parallel attention mechanisms. And we utilize an asymmetric conditional positional encoding scheme which can encode positional information of EEG patches and be easily adapted to the EEG with diverse formats. CBraMod is pre-trained on a very large corpus of EEG through patch-based masked EEG reconstruction. We evaluate CBraMod on up to 10 downstream BCI tasks (12 public datasets). CBraMod achieves the state-of-the-art performance across the wide range of tasks, proving its strong capability and generalizability. The source code is publicly available at \url{https://github.com/wjq-learning/CBraMod}.
☆ Parseval Regularization for Continual Reinforcement Learning
Loss of plasticity, trainability loss, and primacy bias have been identified as issues arising when training deep neural networks on sequences of tasks -- all referring to the increased difficulty in training on new tasks. We propose to use Parseval regularization, which maintains orthogonality of weight matrices, to preserve useful optimization properties and improve training in a continual reinforcement learning setting. We show that it provides significant benefits to RL agents on a suite of gridworld, CARL and MetaWorld tasks. We conduct comprehensive ablations to identify the source of its benefits and investigate the effect of certain metrics associated to network trainability including weight matrix rank, weight norms and policy entropy.
☆ MPSI: Mamba enhancement model for pixel-wise sequential interaction Image Super-Resolution
Single image super-resolution (SR) has long posed a challenge in the field of computer vision. While the advent of deep learning has led to the emergence of numerous methods aimed at tackling this persistent issue, the current methodologies still encounter challenges in modeling long sequence information, leading to limitations in effectively capturing the global pixel interactions. To tackle this challenge and achieve superior SR outcomes, we propose the Mamba pixel-wise sequential interaction network (MPSI), aimed at enhancing the establishment of long-range connections of information, particularly focusing on pixel-wise sequential interaction. We propose the Channel-Mamba Block (CMB) to capture comprehensive pixel interaction information by effectively modeling long sequence information. Moreover, in the existing SR methodologies, there persists the issue of the neglect of features extracted by preceding layers, leading to the loss of valuable feature information. While certain existing models strive to preserve these features, they frequently encounter difficulty in establishing connections across all layers. To overcome this limitation, MPSI introduces the Mamba channel recursion module (MCRM), which maximizes the retention of valuable feature information from early layers, thereby facilitating the acquisition of pixel sequence interaction information from multiple-level layers. Through extensive experimentation, we demonstrate that MPSI outperforms existing super-resolution methods in terms of image reconstruction results, attaining state-of-the-art performance.
☆ Towards Automated Cross-domain Exploratory Data Analysis through Large Language Models SIGMOD 2025
Exploratory data analysis (EDA), coupled with SQL, is essential for data analysts involved in data exploration and analysis. However, data analysts often encounter two primary challenges: (1) the need to craft SQL queries skillfully, and (2) the requirement to generate suitable visualization types that enhance the interpretation of query results. Due to its significance, substantial research efforts have been made to explore different approaches to address these challenges, including leveraging large language models (LLMs). However, existing methods fail to meet real-world data exploration requirements primarily due to (1) complex database schema; (2) unclear user intent; (3) limited cross-domain generalization capability; and (4) insufficient end-to-end text-to-visualization capability. This paper presents TiInsight, an automated SQL-based cross-domain exploratory data analysis system. First, we propose hierarchical data context (i.e., HDC), which leverages LLMs to summarize the contexts related to the database schema, which is crucial for open-world EDA systems to generalize across data domains. Second, the EDA system is divided into four components (i.e., stages): HDC generation, question clarification and decomposition, text-to-SQL generation (i.e., TiSQL), and data visualization (i.e., TiChart). Finally, we implemented an end-to-end EDA system with a user-friendly GUI interface in the production environment at PingCAP. We have also open-sourced all APIs of TiInsight to facilitate research within the EDA community. Through extensive evaluations by a real-world user study, we demonstrate that TiInsight offers remarkable performance compared to human experts. Specifically, TiSQL achieves an execution accuracy of 86.3% on the Spider dataset using GPT-4. It also demonstrates state-of-the-art performance on the Bird dataset.
comment: 14 pages, 10 figures. Submitted to SIGMOD 2025
☆ IntellectSeeker: A Personalized Literature Management System with the Probabilistic Model and Large Language Model
Faced with the burgeoning volume of academic literature, researchers often need help with uncertain article quality and mismatches in term searches using traditional academic engines. We introduce IntellectSeeker, an innovative and personalized intelligent academic literature management platform to address these challenges. This platform integrates a Large Language Model (LLM)--based semantic enhancement bot with a sophisticated probability model to personalize and streamline literature searches. We adopted the GPT-3.5-turbo model to transform everyday language into professional academic terms across various scenarios using multiple rounds of few-shot learning. This adaptation mainly benefits academic newcomers, effectively bridging the gap between general inquiries and academic terminology. The probabilistic model intelligently filters academic articles to align closely with the specific interests of users, which are derived from explicit needs and behavioral patterns. Moreover, IntellectSeeker incorporates an advanced recommendation system and text compression tools. These features enable intelligent article recommendations based on user interactions and present search results through concise one-line summaries and innovative word cloud visualizations, significantly enhancing research efficiency and user experience. IntellectSeeker offers academic researchers a highly customizable literature management solution with exceptional search precision and matching capabilities. The code can be found here: https://github.com/LuckyBian/ISY5001
☆ EDiT: A Local-SGD-Based Efficient Distributed Training Method for Large Language Models
Distributed training methods are crucial for large language models (LLMs). However, existing distributed training methods often suffer from communication bottlenecks, stragglers, and limited elasticity. Local SGD methods have been proposed to address these issues, but their effectiveness remains limited to small-scale training due to additional memory overhead and lack of concerns on efficiency and stability. To tackle these issues, we propose EDiT, an innovative Efficient Distributed Training method that combines a tailored Local SGD approach with model sharding techniques to enhance large-scale training efficiency. EDiT performs layer-wise parameter synchronization during forward pass, reducing communication and memory overhead and enabling the overlap of computation and communication. Besides, EDiT employs a pseudo gradient penalty strategy to suppress loss spikes, which ensures training stability and improve performance. Additionally, we introduce A-EDiT, a fully asynchronous variant of EDiT that accommodates heterogeneous clusters. Building on EDiT/A-EDiT, we conduct a series of experiments to validate large-scale asynchronous training for LLMs, accompanied by comprehensive analyses. Experimental results demonstrate the superior performance of EDiT/A-EDiT, establishing them as robust solutions for distributed LLM training in diverse computational ecosystems.
comment: 22 pages, 10 figures, 7 tables
☆ MAPLE: A Framework for Active Preference Learning Guided by Large Language Models
The advent of large language models (LLMs) has sparked significant interest in using natural language for preference learning. However, existing methods often suffer from high computational burdens, taxing human supervision, and lack of interpretability. To address these issues, we introduce MAPLE, a framework for large language model-guided Bayesian active preference learning. MAPLE leverages LLMs to model the distribution over preference functions, conditioning it on both natural language feedback and conventional preference learning feedback, such as pairwise trajectory rankings. MAPLE also employs active learning to systematically reduce uncertainty in this distribution and incorporates a language-conditioned active query selection mechanism to identify informative and easy-to-answer queries, thus reducing human burden. We evaluate MAPLE's sample efficiency and preference inference quality across two benchmarks, including a real-world vehicle route planning benchmark using OpenStreetMap data. Our results demonstrate that MAPLE accelerates the learning process and effectively improves humans' ability to answer queries.
☆ A Review on the Applications of Transformer-based language models for Nucleotide Sequence Analysis
In recent times, Transformer-based language models are making quite an impact in the field of natural language processing. As relevant parallels can be drawn between biological sequences and natural languages, the models used in NLP can be easily extended and adapted for various applications in bioinformatics. In this regard, this paper introduces the major developments of Transformer-based models in the recent past in the context of nucleotide sequences. We have reviewed and analysed a large number of application-based papers on this subject, giving evidence of the main characterizing features and to different approaches that may be adopted to customize such powerful computational machines. We have also provided a structured description of the functioning of Transformers, that may enable even first time users to grab the essence of such complex architectures. We believe this review will help the scientific community in understanding the various applications of Transformer-based language models to nucleotide sequences. This work will motivate the readers to build on these methodologies to tackle also various other problems in the field of bioinformatics.
☆ Modifying AI, Enhancing Essays: How Active Engagement with Generative AI Boosts Writing Quality
Students are increasingly relying on Generative AI (GAI) to support their writing-a key pedagogical practice in education. In GAI-assisted writing, students can delegate core cognitive tasks (e.g., generating ideas and turning them into sentences) to GAI while still producing high-quality essays. This creates new challenges for teachers in assessing and supporting student learning, as they often lack insight into whether students are engaging in meaningful cognitive processes during writing or how much of the essay's quality can be attributed to those processes. This study aimed to help teachers better assess and support student learning in GAI-assisted writing by examining how different writing behaviors, especially those indicative of meaningful learning versus those that are not, impact essay quality. Using a dataset of 1,445 GAI-assisted writing sessions, we applied the cutting-edge method, X-Learner, to quantify the causal impact of three GAI-assisted writing behavioral patterns (i.e., seeking suggestions but not accepting them, seeking suggestions and accepting them as they are, and seeking suggestions and accepting them with modification) on four measures of essay quality (i.e., lexical sophistication, syntactic complexity, text cohesion, and linguistic bias). Our analysis showed that writers who frequently modified GAI-generated text-suggesting active engagement in higher-order cognitive processes-consistently improved the quality of their essays in terms of lexical sophistication, syntactic complexity, and text cohesion. In contrast, those who often accepted GAI-generated text without changes, primarily engaging in lower-order processes, saw a decrease in essay quality. Additionally, while human writers tend to introduce linguistic bias when writing independently, incorporating GAI-generated text-even without modification-can help mitigate this bias.
☆ Hierarchical Split Federated Learning: Convergence Analysis and System Optimization
As AI models expand in size, it has become increasingly challenging to deploy federated learning (FL) on resource-constrained edge devices. To tackle this issue, split federated learning (SFL) has emerged as an FL framework with reduced workload on edge devices via model splitting; it has received extensive attention from the research community in recent years. Nevertheless, most prior works on SFL focus only on a two-tier architecture without harnessing multi-tier cloudedge computing resources. In this paper, we intend to analyze and optimize the learning performance of SFL under multi-tier systems. Specifically, we propose the hierarchical SFL (HSFL) framework and derive its convergence bound. Based on the theoretical results, we formulate a joint optimization problem for model splitting (MS) and model aggregation (MA). To solve this rather hard problem, we then decompose it into MS and MA subproblems that can be solved via an iterative descending algorithm. Simulation results demonstrate that the tailored algorithm can effectively optimize MS and MA for SFL within virtually any multi-tier system.
comment: 15 pages, 9 figures
☆ Graph Neural Networks Are More Than Filters: Revisiting and Benchmarking from A Spectral Perspective
Graph Neural Networks (GNNs) have achieved remarkable success in various graph-based learning tasks. While their performance is often attributed to the powerful neighborhood aggregation mechanism, recent studies suggest that other components such as non-linear layers may also significantly affecting how GNNs process the input graph data in the spectral domain. Such evidence challenges the prevalent opinion that neighborhood aggregation mechanisms dominate the behavioral characteristics of GNNs in the spectral domain. To demystify such a conflict, this paper introduces a comprehensive benchmark to measure and evaluate GNNs' capability in capturing and leveraging the information encoded in different frequency components of the input graph data. Specifically, we first conduct an exploratory study demonstrating that GNNs can flexibly yield outputs with diverse frequency components even when certain frequencies are absent or filtered out from the input graph data. We then formulate a novel research problem of measuring and benchmarking the performance of GNNs from a spectral perspective. To take an initial step towards a comprehensive benchmark, we design an evaluation protocol supported by comprehensive theoretical analysis. Finally, we introduce a comprehensive benchmark on real-world datasets, revealing insights that challenge prevalent opinions from a spectral perspective. We believe that our findings will open new avenues for future advancements in this area. Our implementations can be found at: https://github.com/yushundong/Spectral-benchmark.
☆ Monte Carlo Tree Search based Space Transfer for Black-box Optimization NeurIPS 2024
Bayesian optimization (BO) is a popular method for computationally expensive black-box optimization. However, traditional BO methods need to solve new problems from scratch, leading to slow convergence. Recent studies try to extend BO to a transfer learning setup to speed up the optimization, where search space transfer is one of the most promising approaches and has shown impressive performance on many tasks. However, existing search space transfer methods either lack an adaptive mechanism or are not flexible enough, making it difficult to efficiently identify promising search space during the optimization process. In this paper, we propose a search space transfer learning method based on Monte Carlo tree search (MCTS), called MCTS-transfer, to iteratively divide, select, and optimize in a learned subspace. MCTS-transfer can not only provide a well-performing search space for warm-start but also adaptively identify and leverage the information of similar source tasks to reconstruct the search space during the optimization process. Experiments on synthetic functions, real-world problems, Design-Bench and hyper-parameter optimization show that MCTS-transfer can demonstrate superior performance compared to other search space transfer methods under different settings. Our code is available at \url{https://github.com/lamda-bbo/mcts-transfer}.
comment: NeurIPS 2024 Spotlight
☆ Exploring What Why and How: A Multifaceted Benchmark for Causation Understanding of Video Anomaly
Recent advancements in video anomaly understanding (VAU) have opened the door to groundbreaking applications in various fields, such as traffic monitoring and industrial automation. While the current benchmarks in VAU predominantly emphasize the detection and localization of anomalies. Here, we endeavor to delve deeper into the practical aspects of VAU by addressing the essential questions: "what anomaly occurred?", "why did it happen?", and "how severe is this abnormal event?". In pursuit of these answers, we introduce a comprehensive benchmark for Exploring the Causation of Video Anomalies (ECVA). Our benchmark is meticulously designed, with each video accompanied by detailed human annotations. Specifically, each instance of our ECVA involves three sets of human annotations to indicate "what", "why" and "how" of an anomaly, including 1) anomaly type, start and end times, and event descriptions, 2) natural language explanations for the cause of an anomaly, and 3) free text reflecting the effect of the abnormality. Building upon this foundation, we propose a novel prompt-based methodology that serves as a baseline for tackling the intricate challenges posed by ECVA. We utilize "hard prompt" to guide the model to focus on the critical parts related to video anomaly segments, and "soft prompt" to establish temporal and spatial relationships within these anomaly segments. Furthermore, we propose AnomEval, a specialized evaluation metric crafted to align closely with human judgment criteria for ECVA. This metric leverages the unique features of the ECVA dataset to provide a more comprehensive and reliable assessment of various video large language models. We demonstrate the efficacy of our approach through rigorous experimental analysis and delineate possible avenues for further investigation into the comprehension of video anomaly causation.
comment: Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence. arXiv admin note: substantial text overlap with arXiv:2405.00181
☆ An Enhancement of CNN Algorithm for Rice Leaf Disease Image Classification in Mobile Applications
This study focuses on enhancing rice leaf disease image classification algorithms, which have traditionally relied on Convolutional Neural Network (CNN) models. We employed transfer learning with MobileViTV2_050 using ImageNet-1k weights, a lightweight model that integrates CNN's local feature extraction with Vision Transformers' global context learning through a separable self-attention mechanism. Our approach resulted in a significant 15.66% improvement in classification accuracy for MobileViTV2_050-A, our first enhanced model trained on the baseline dataset, achieving 93.14%. Furthermore, MobileViTV2_050-B, our second enhanced model trained on a broader rice leaf dataset, demonstrated a 22.12% improvement, reaching 99.6% test accuracy. Additionally, MobileViTV2-A attained an F1-score of 93% across four rice labels and a Receiver Operating Characteristic (ROC) curve ranging from 87% to 97%. In terms of resource consumption, our enhanced models reduced the total parameters of the baseline CNN model by up to 92.50%, from 14 million to 1.1 million. These results indicate that MobileViTV2_050 not only improves computational efficiency through its separable self-attention mechanism but also enhances global context learning. Consequently, it offers a lightweight and robust solution suitable for mobile deployment, advancing the interpretability and practicality of models in precision agriculture.
comment: Presented at 46th World Conference on Applied Science, Engineering & Technology (WCASET) from Institute for Educational Research and Publication (IFERP)
☆ Post-Training Statistical Calibration for Higher Activation Sparsity SP
We present Statistical Calibrated Activation Pruning (SCAP), a post-training activation pruning framework that (1) generalizes sparsification by input activations of Fully-Connected layers for generic and flexible application across Transformers, and (2) features a simple Mode-Centering technique to pre-calibrate activation distributions for maximizing post-training sparsity. Our results demonstrate robust Pareto efficiency compared to prior methods, translating to a 1.5x additional LLM decoding speedup against CATS at iso model quality. SCAP effectiveness is empirically verified across a wide range of models, including recent Transformer Decoders, MoE, Mamba2, Encoding Transformer, and pre-quantized models, highlighting its practicality and scalability. The code is available at: https://github.com/IntelLabs/SCAP.
comment: ENLSP-IV NeurIPS Workshop 2024
☆ Reinforcement Learning Policy as Macro Regulator Rather than Macro Placer NeurIPS 2024
In modern chip design, placement aims at placing millions of circuit modules, which is an essential step that significantly influences power, performance, and area (PPA) metrics. Recently, reinforcement learning (RL) has emerged as a promising technique for improving placement quality, especially macro placement. However, current RL-based placement methods suffer from long training times, low generalization ability, and inability to guarantee PPA results. A key issue lies in the problem formulation, i.e., using RL to place from scratch, which results in limits useful information and inaccurate rewards during the training process. In this work, we propose an approach that utilizes RL for the refinement stage, which allows the RL policy to learn how to adjust existing placement layouts, thereby receiving sufficient information for the policy to act and obtain relatively dense and precise rewards. Additionally, we introduce the concept of regularity during training, which is considered an important metric in the chip design industry but is often overlooked in current RL placement methods. We evaluate our approach on the ISPD 2005 and ICCAD 2015 benchmark, comparing the global half-perimeter wirelength and regularity of our proposed method against several competitive approaches. Besides, we test the PPA performance using commercial software, showing that RL as a regulator can achieve significant PPA improvements. Our RL regulator can fine-tune placements from any method and enhance their quality. Our work opens up new possibilities for the application of RL in placement, providing a more effective and efficient approach to optimizing chip design. Our code is available at \url{https://github.com/lamda-bbo/macro-regulator}.
comment: NeurIPS 2024
☆ A Method for Evaluating Hyperparameter Sensitivity in Reinforcement Learning
The performance of modern reinforcement learning algorithms critically relies on tuning ever-increasing numbers of hyperparameters. Often, small changes in a hyperparameter can lead to drastic changes in performance, and different environments require very different hyperparameter settings to achieve state-of-the-art performance reported in the literature. We currently lack a scalable and widely accepted approach to characterizing these complex interactions. This work proposes a new empirical methodology for studying, comparing, and quantifying the sensitivity of an algorithm's performance to hyperparameter tuning for a given set of environments. We then demonstrate the utility of this methodology by assessing the hyperparameter sensitivity of several commonly used normalization variants of PPO. The results suggest that several algorithmic performance improvements may, in fact, be a result of an increased reliance on hyperparameter tuning.
☆ Fast Occupancy Network
Occupancy Network has recently attracted much attention in autonomous driving. Instead of monocular 3D detection and recent bird's eye view(BEV) models predicting 3D bounding box of obstacles, Occupancy Network predicts the category of voxel in specified 3D space around the ego vehicle via transforming 3D detection task into 3D voxel segmentation task, which has much superiority in tackling category outlier obstacles and providing fine-grained 3D representation. However, existing methods usually require huge computation resources than previous methods, which hinder the Occupancy Network solution applying in intelligent driving systems. To address this problem, we make an analysis of the bottleneck of Occupancy Network inference cost, and present a simple and fast Occupancy Network model, which adopts a deformable 2D convolutional layer to lift BEV feature to 3D voxel feature and presents an efficient voxel feature pyramid network (FPN) module to improve performance with few computational cost. Further, we present a cost-free 2D segmentation branch in perspective view after feature extractors for Occupancy Network during inference phase to improve accuracy. Experimental results demonstrate that our method consistently outperforms existing methods in both accuracy and inference speed, which surpasses recent state-of-the-art (SOTA) OCCNet by 1.7% with ResNet50 backbone with about 3X inference speedup. Furthermore, our method can be easily applied to existing BEV models to transform them into Occupancy Network models.
comment: 10 pages, 5 figures,
☆ MM-PoE: Multiple Choice Reasoning via. Process of Elimination using Multi-Modal Models
This paper introduces Multiple Choice Reasoning via. Process of Elimination using Multi-Modal models, herein referred to as Multi-Modal Process of Elimination (MM-PoE). This novel methodology is engineered to augment the efficacy of Vision-Language Models (VLMs) in multiple-choice visual reasoning tasks. Diverging from conventional approaches that evaluate each option independently, MM-PoE employs a dual-step scoring paradigm that initially identifies and excludes implausible choices, subsequently concentrating on the most probable remaining options. This method emulates human test-taking strategies, where individuals typically eliminate clearly incorrect answers prior to selecting the optimal response. Our empirical evaluations, conducted across three benchmark datasets, reveal that MM-PoE significantly improves both zero-shot and few-shot performance of contemporary state-of-the-art VLMs. Critically, this approach not only broadens the application of the elimination process to multi-modal contexts but also allows few-shot experiments, thereby addressing two principal limitations concerning usage of PoE only in zero-shot settings and only with a language-only framework. As a result, MM-PoE not only refines the reasoning capabilities of VLMs but also broadens their applicability to complex visual question-answering scenarios. All code and documentation supporting our work are available at https://pypi.org/project/mm-poe/, enabling researchers and practitioners to easily integrate and further develop these techniques.
☆ MIT-10M: A Large Scale Parallel Corpus of Multilingual Image Translation COLING 2025
Image Translation (IT) holds immense potential across diverse domains, enabling the translation of textual content within images into various languages. However, existing datasets often suffer from limitations in scale, diversity, and quality, hindering the development and evaluation of IT models. To address this issue, we introduce MIT-10M, a large-scale parallel corpus of multilingual image translation with over 10M image-text pairs derived from real-world data, which has undergone extensive data cleaning and multilingual translation validation. It contains 840K images in three sizes, 28 categories, tasks with three levels of difficulty and 14 languages image-text pairs, which is a considerable improvement on existing datasets. We conduct extensive experiments to evaluate and train models on MIT-10M. The experimental results clearly indicate that our dataset has higher adaptability when it comes to evaluating the performance of the models in tackling challenging and complex image translation tasks in the real world. Moreover, the performance of the model fine-tuned with MIT-10M has tripled compared to the baseline model, further confirming its superiority.
comment: Accepted in COLING 2025
☆ Political Actor Agent: Simulating Legislative System for Roll Call Votes Prediction with Large Language Models AAAI 2025
Predicting roll call votes through modeling political actors has emerged as a focus in quantitative political science and computer science. Widely used embedding-based methods generate vectors for legislators from diverse data sets to predict legislative behaviors. However, these methods often contend with challenges such as the need for manually predefined features, reliance on extensive training data, and a lack of interpretability. Achieving more interpretable predictions under flexible conditions remains an unresolved issue. This paper introduces the Political Actor Agent (PAA), a novel agent-based framework that utilizes Large Language Models to overcome these limitations. By employing role-playing architectures and simulating legislative system, PAA provides a scalable and interpretable paradigm for predicting roll-call votes. Our approach not only enhances the accuracy of predictions but also offers multi-view, human-understandable decision reasoning, providing new insights into political actor behaviors. We conducted comprehensive experiments using voting records from the 117-118th U.S. House of Representatives, validating the superior performance and interpretability of PAA. This study not only demonstrates PAA's effectiveness but also its potential in political science research.
comment: Accepted at AAAI 2025
☆ Deep Learning-Enhanced Preconditioning for Efficient Conjugate Gradient Solvers in Large-Scale PDE Systems
Preconditioning techniques are crucial for enhancing the efficiency of solving large-scale linear equation systems that arise from partial differential equation (PDE) discretization. These techniques, such as Incomplete Cholesky factorization (IC) and data-driven neural network methods, accelerate the convergence of iterative solvers like Conjugate Gradient (CG) by approximating the original matrices. This paper introduces a novel approach that integrates Graph Neural Network (GNN) with traditional IC, addressing the shortcomings of direct generation methods based on GNN and achieving significant improvements in computational efficiency and scalability. Experimental results demonstrate an average reduction in iteration counts by 24.8% compared to IC and a two-order-of-magnitude increase in training scale compared to previous methods. A three-dimensional static structural analysis utilizing finite element methods was validated on training sparse matrices of up to 5 million dimensions and inference scales of up to 10 million. Furthermore, the approach demon-strates robust generalization capabilities across scales, facilitating the effective acceleration of CG solvers for large-scale linear equations using small-scale data on modest hardware. The method's robustness and scalability make it a practical solution for computational science.
☆ A Review of Human Emotion Synthesis Based on Generative Technology
Human emotion synthesis is a crucial aspect of affective computing. It involves using computational methods to mimic and convey human emotions through various modalities, with the goal of enabling more natural and effective human-computer interactions. Recent advancements in generative models, such as Autoencoders, Generative Adversarial Networks, Diffusion Models, Large Language Models, and Sequence-to-Sequence Models, have significantly contributed to the development of this field. However, there is a notable lack of comprehensive reviews in this field. To address this problem, this paper aims to address this gap by providing a thorough and systematic overview of recent advancements in human emotion synthesis based on generative models. Specifically, this review will first present the review methodology, the emotion models involved, the mathematical principles of generative models, and the datasets used. Then, the review covers the application of different generative models to emotion synthesis based on a variety of modalities, including facial images, speech, and text. It also examines mainstream evaluation metrics. Additionally, the review presents some major findings and suggests future research directions, providing a comprehensive understanding of the role of generative technology in the nuanced domain of emotion synthesis.
comment: 25 pages, 10 figures
☆ On Evaluating the Durability of Safeguards for Open-Weight LLMs
Stakeholders -- from model developers to policymakers -- seek to minimize the dual-use risks of large language models (LLMs). An open challenge to this goal is whether technical safeguards can impede the misuse of LLMs, even when models are customizable via fine-tuning or when model weights are fully open. In response, several recent studies have proposed methods to produce durable LLM safeguards for open-weight LLMs that can withstand adversarial modifications of the model's weights via fine-tuning. This holds the promise of raising adversaries' costs even under strong threat models where adversaries can directly fine-tune model weights. However, in this paper, we urge for more careful characterization of the limits of these approaches. Through several case studies, we demonstrate that even evaluating these defenses is exceedingly difficult and can easily mislead audiences into thinking that safeguards are more durable than they really are. We draw lessons from the evaluation pitfalls that we identify and suggest future research carefully cabin claims to more constrained, well-defined, and rigorously examined threat models, which can provide more useful and candid assessments to stakeholders.
☆ QAPyramid: Fine-grained Evaluation of Content Selection for Text Summarization
How to properly conduct human evaluations for text summarization is a longstanding challenge. The Pyramid human evaluation protocol, which assesses content selection by breaking the reference summary into sub-units and verifying their presence in the system summary, has been widely adopted. However, it suffers from a lack of systematicity in the definition and granularity of the sub-units. We address these problems by proposing QAPyramid, which decomposes each reference summary into finer-grained question-answer (QA) pairs according to the QA-SRL framework. We collect QA-SRL annotations for reference summaries from CNN/DM and evaluate 10 summarization systems, resulting in 8.9K QA-level annotations. We show that, compared to Pyramid, QAPyramid provides more systematic and fine-grained content selection evaluation while maintaining high inter-annotator agreement without needing expert annotations. Furthermore, we propose metrics that automate the evaluation pipeline and achieve higher correlations with QAPyramid than other widely adopted metrics, allowing future work to accurately and efficiently benchmark summarization systems.
comment: The first two authors contributed equally. Code: https://github.com/ZhangShiyue/QAPyramid
☆ Access Point Deployment for Localizing Accuracy and User Rate in Cell-Free Systems
Evolving next-generation mobile networks is designed to provide ubiquitous coverage and networked sensing. With utility of multi-view sensing and multi-node joint transmission, cell-free is a promising technique to realize this prospect. This paper aims to tackle the problem of access point (AP) deployment in cell-free systems to balance the sensing accuracy and user rate. By merging the D-optimality with Euclidean criterion, a novel integrated metric is proposed to be the objective function for both max-sum and max-min problems, which respectively guarantee the overall and lowest performance in multi-user communication and target tracking scenario. To solve the corresponding high dimensional non-convex multi-objective problem, the Soft actor-critic (SAC) is utilized to avoid risk of local optimal result. Numerical results demonstrate that proposed SAC-based APs deployment method achieves $20\%$ of overall performance and $120\%$ of lowest performance.
comment: Presented at MobiCom 2024
☆ Sequential Controlled Langevin Diffusions
An effective approach for sampling from unnormalized densities is based on the idea of gradually transporting samples from an easy prior to the complicated target distribution. Two popular methods are (1) Sequential Monte Carlo (SMC), where the transport is performed through successive annealed densities via prescribed Markov chains and resampling steps, and (2) recently developed diffusion-based sampling methods, where a learned dynamical transport is used. Despite the common goal, both approaches have different, often complementary, advantages and drawbacks. The resampling steps in SMC allow focusing on promising regions of the space, often leading to robust performance. While the algorithm enjoys asymptotic guarantees, the lack of flexible, learnable transitions can lead to slow convergence. On the other hand, diffusion-based samplers are learned and can potentially better adapt themselves to the target at hand, yet often suffer from training instabilities. In this work, we present a principled framework for combining SMC with diffusion-based samplers by viewing both methods in continuous time and considering measures on path space. This culminates in the new Sequential Controlled Langevin Diffusion (SCLD) sampling method, which is able to utilize the benefits of both methods and reaches improved performance on multiple benchmark problems, in many cases using only 10% of the training budget of previous diffusion-based samplers.
☆ EvRepSL: Event-Stream Representation via Self-Supervised Learning for Event-Based Vision
Event-stream representation is the first step for many computer vision tasks using event cameras. It converts the asynchronous event-streams into a formatted structure so that conventional machine learning models can be applied easily. However, most of the state-of-the-art event-stream representations are manually designed and the quality of these representations cannot be guaranteed due to the noisy nature of event-streams. In this paper, we introduce a data-driven approach aiming at enhancing the quality of event-stream representations. Our approach commences with the introduction of a new event-stream representation based on spatial-temporal statistics, denoted as EvRep. Subsequently, we theoretically derive the intrinsic relationship between asynchronous event-streams and synchronous video frames. Building upon this theoretical relationship, we train a representation generator, RepGen, in a self-supervised learning manner accepting EvRep as input. Finally, the event-streams are converted to high-quality representations, termed as EvRepSL, by going through the learned RepGen (without the need of fine-tuning or retraining). Our methodology is rigorously validated through extensive evaluations on a variety of mainstream event-based classification and optical flow datasets (captured with various types of event cameras). The experimental results highlight not only our approach's superior performance over existing event-stream representations but also its versatility, being agnostic to different event cameras and tasks.
comment: Published on IEEE Transactions on Image Processing
☆ Defensive Dual Masking for Robust Adversarial Defense
The field of textual adversarial defenses has gained considerable attention in recent years due to the increasing vulnerability of natural language processing (NLP) models to adversarial attacks, which exploit subtle perturbations in input text to deceive models. This paper introduces the Defensive Dual Masking (DDM) algorithm, a novel approach designed to enhance model robustness against such attacks. DDM utilizes a unique adversarial training strategy where [MASK] tokens are strategically inserted into training samples to prepare the model to handle adversarial perturbations more effectively. During inference, potentially adversarial tokens are dynamically replaced with [MASK] tokens to neutralize potential threats while preserving the core semantics of the input. The theoretical foundation of our approach is explored, demonstrating how the selective masking mechanism strengthens the model's ability to identify and mitigate adversarial manipulations. Our empirical evaluation across a diverse set of benchmark datasets and attack mechanisms consistently shows that DDM outperforms state-of-the-art defense techniques, improving model accuracy and robustness. Moreover, when applied to Large Language Models (LLMs), DDM also enhances their resilience to adversarial attacks, providing a scalable defense mechanism for large-scale NLP applications.
comment: First version
☆ The Mirage of Artificial Intelligence Terms of Use Restrictions
Artificial intelligence (AI) model creators commonly attach restrictive terms of use to both their models and their outputs. These terms typically prohibit activities ranging from creating competing AI models to spreading disinformation. Often taken at face value, these terms are positioned by companies as key enforceable tools for preventing misuse, particularly in policy dialogs. But are these terms truly meaningful? There are myriad examples where these broad terms are regularly and repeatedly violated. Yet except for some account suspensions on platforms, no model creator has actually tried to enforce these terms with monetary penalties or injunctive relief. This is likely for good reason: we think that the legal enforceability of these licenses is questionable. This Article systematically assesses of the enforceability of AI model terms of use and offers three contributions. First, we pinpoint a key problem: the artifacts that they protect, namely model weights and model outputs, are largely not copyrightable, making it unclear whether there is even anything to be licensed. Second, we examine the problems this creates for other enforcement. Recent doctrinal trends in copyright preemption may further undermine state-law claims, while other legal frameworks like the DMCA and CFAA offer limited recourse. Anti-competitive provisions likely fare even worse than responsible use provisions. Third, we provide recommendations to policymakers. There are compelling reasons for many provisions to be unenforceable: they chill good faith research, constrain competition, and create quasi-copyright ownership where none should exist. There are, of course, downsides: model creators have fewer tools to prevent harmful misuse. But we think the better approach is for statutory provisions, not private fiat, to distinguish between good and bad uses of AI, restricting the latter.
comment: Forthcoming Indiana Law Journal
☆ Where Common Knowledge Cannot Be Formed, Common Belief Can -- Planning with Multi-Agent Belief Using Group Justified Perspectives
Epistemic planning is the sub-field of AI planning that focuses on changing knowledge and belief. It is important in both multi-agent domains where agents need to have knowledge/belief regarding the environment, but also the beliefs of other agents, including nested beliefs. When modeling knowledge in multi-agent settings, many models face an exponential growth challenge in terms of nested depth. A contemporary method, known as Planning with Perspectives (PWP), addresses these challenges through the use of perspectives and set operations for knowledge. The JP model defines that an agent's belief is justified if and only if the agent has seen evidence that this belief was true in the past and has not seen evidence to suggest that this has changed. The current paper extends the JP model to handle \emph{group belief}, including distributed belief and common belief. We call this the Group Justified Perspective (GJP) model. Using experimental problems crafted by adapting well-known benchmarks to a group setting, we show the efficiency and expressiveness of our GJP model at handling planning problems that cannot be handled by other epistemic planning tools.
comment: 10 pages, including appendix and reference
♻ ☆ XRZoo: A Large-Scale and Versatile Dataset of Extended Reality (XR) Applications
The rapid advancement of Extended Reality (XR, encompassing AR, MR, and VR) and spatial computing technologies forms a foundational layer for the emerging Metaverse, enabling innovative applications across healthcare, education, manufacturing, and entertainment. However, research in this area is often limited by the lack of large, representative, and highquality application datasets that can support empirical studies and the development of new approaches benefiting XR software processes. In this paper, we introduce XRZoo, a comprehensive and curated dataset of XR applications designed to bridge this gap. XRZoo contains 12,528 free XR applications, spanning nine app stores, across all XR techniques (i.e., AR, MR, and VR) and use cases, with detailed metadata on key aspects such as application descriptions, application categories, release dates, user review numbers, and hardware specifications, etc. By making XRZoo publicly available, we aim to foster reproducible XR software engineering and security research, enable cross-disciplinary investigations, and also support the development of advanced XR systems by providing examples to developers. Our dataset serves as a valuable resource for researchers and practitioners interested in improving the scalability, usability, and effectiveness of XR applications. XRZoo will be released and actively maintained.
♻ ☆ Impact of Leakage on Data Harmonization in Machine Learning Pipelines in Class Imbalance Across Sites
Machine learning (ML) models benefit from large datasets. Collecting data in biomedical domains is costly and challenging, hence, combining datasets has become a common practice. However, datasets obtained under different conditions could present undesired site-specific variability. Data harmonization methods aim to remove site-specific variance while retaining biologically relevant information. This study evaluates the effectiveness of popularly used ComBat-based methods for harmonizing data in scenarios where the class balance is not equal across sites. We find that these methods struggle with data leakage issues. To overcome this problem, we propose a novel approach PrettYharmonize, designed to harmonize data by pretending the target labels. We validate our approach using controlled datasets designed to benchmark the utility of harmonization. Finally, using real-world MRI and clinical data, we compare leakage-prone methods with PrettYharmonize and show that it achieves comparable performance while avoiding data leakage, particularly in site-target-dependence scenarios.
♻ ☆ Beyond Retrieval: Generating Narratives in Conversational Recommender Systems
The recent advances in Large Language Model's generation and reasoning capabilities present an opportunity to develop truly conversational recommendation systems. However, effectively integrating recommender system knowledge into LLMs for natural language generation which is tailored towards recommendation tasks remains a challenge. This paper addresses this challenge by making two key contributions. First, we introduce a new dataset (REGEN) for natural language generation tasks in conversational recommendations. REGEN (Reviews Enhanced with GEnerative Narratives) extends the Amazon Product Reviews dataset with rich user narratives, including personalized explanations of product preferences, product endorsements for recommended items, and summaries of user purchase history. REGEN is made publicly available to facilitate further research. Furthermore, we establish benchmarks using well-known generative metrics, and perform an automated evaluation of the new dataset using a rater LLM. Second, the paper introduces a fusion architecture (CF model with an LLM) which serves as a baseline for REGEN. And to the best of our knowledge, represents the first attempt to analyze the capabilities of LLMs in understanding recommender signals and generating rich narratives. We demonstrate that LLMs can effectively learn from simple fusion architectures utilizing interaction-based CF embeddings, and this can be further enhanced using the metadata and personalization data associated with items. Our experiments show that combining CF and content embeddings leads to improvements of 4-12% in key language metrics compared to using either type of embedding individually. We also provide an analysis to interpret how CF and content embeddings contribute to this new generative task.
♻ ☆ The BrowserGym Ecosystem for Web Agent Research
The BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents, particularly those leveraging automation and Large Language Models (LLMs) for web interaction tasks. Many existing benchmarks suffer from fragmentation and inconsistent evaluation methodologies, making it challenging to achieve reliable comparisons and reproducible results. BrowserGym aims to solve this by providing a unified, gym-like environment with well-defined observation and action spaces, facilitating standardized evaluation across diverse benchmarks. Combined with AgentLab, a complementary framework that aids in agent creation, testing, and analysis, BrowserGym offers flexibility for integrating new benchmarks while ensuring consistent evaluation and comprehensive experiment management. This standardized approach seeks to reduce the time and complexity of developing web agents, supporting more reliable comparisons and facilitating in-depth analysis of agent behaviors, and could result in more adaptable, capable agents, ultimately accelerating innovation in LLM-driven automation. As a supporting evidence, we conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across all benchmarks currently available in BrowserGym. Among other findings, our results highlight a large discrepancy between OpenAI and Anthropic's latests models, with Claude-3.5-Sonnet leading the way on almost all benchmarks, except on vision-related tasks where GPT-4o is superior. Despite these advancements, our results emphasize that building robust and efficient web agents remains a significant challenge, due to the inherent complexity of real-world web environments and the limitations of current models.
♻ ☆ Do graph neural network states contain graph properties?
Deep neural networks (DNNs) achieve state-of-the-art performance on many tasks, but this often requires increasingly larger model sizes, which in turn leads to more complex internal representations. Explainability techniques (XAI) have made remarkable progress in the interpretability of ML models. However, the non-relational nature of Graph neural networks (GNNs) make it difficult to reuse already existing XAI methods. While other works have focused on instance-based explanation methods for GNNs, very few have investigated model-based methods and, to our knowledge, none have tried to probe the embedding of the GNNs for well-known structural graph properties. In this paper we present a model agnostic explainability pipeline for GNNs employing diagnostic classifiers. This pipeline aims to probe and interpret the learned representations in GNNs across various architectures and datasets, refining our understanding and trust in these models.
comment: 10 pages, 22 figures, conference
♻ ☆ An Entailment Tree Generation Approach for Multimodal Multi-Hop Question Answering with Mixture-of-Experts and Iterative Feedback Mechanism
With the rise of large-scale language models (LLMs), it is currently popular and effective to convert multimodal information into text descriptions for multimodal multi-hop question answering. However, we argue that the current methods of multi-modal multi-hop question answering still mainly face two challenges: 1) The retrieved evidence containing a large amount of redundant information, inevitably leads to a significant drop in performance due to irrelevant information misleading the prediction. 2) The reasoning process without interpretable reasoning steps makes the model difficult to discover the logical errors for handling complex questions. To solve these problems, we propose a unified LLMs-based approach but without heavily relying on them due to the LLM's potential errors, and innovatively treat multimodal multi-hop question answering as a joint entailment tree generation and question answering problem. Specifically, we design a multi-task learning framework with a focus on facilitating common knowledge sharing across interpretability and prediction tasks while preventing task-specific errors from interfering with each other via mixture of experts. Afterward, we design an iterative feedback mechanism to further enhance both tasks by feeding back the results of the joint training to the LLM for regenerating entailment trees, aiming to iteratively refine the potential answer. Notably, our method has won the first place in the official leaderboard of WebQA (since April 10, 2024), and achieves competitive results on MultimodalQA.
comment: Erratum: We identified an error in the calculation of the F1 score in table 4 reported in a previous version of this work. The performance of the new result is better than the previous one. The corrected values are included in this updated version of the paper. These changes do not alter the primary conclusions of our research
♻ ☆ AdaSociety: An Adaptive Environment with Social Structures for Multi-Agent Decision-Making NeurIPS
Traditional interactive environments limit agents' intelligence growth with fixed tasks. Recently, single-agent environments address this by generating new tasks based on agent actions, enhancing task diversity. We consider the decision-making problem in multi-agent settings, where tasks are further influenced by social connections, affecting rewards and information access. However, existing multi-agent environments lack a combination of adaptive physical surroundings and social connections, hindering the learning of intelligent behaviors. To address this, we introduce AdaSociety, a customizable multi-agent environment featuring expanding state and action spaces, alongside explicit and alterable social structures. As agents progress, the environment adaptively generates new tasks with social structures for agents to undertake. In AdaSociety, we develop three mini-games showcasing distinct social structures and tasks. Initial results demonstrate that specific social structures can promote both individual and collective benefits, though current reinforcement learning and LLM-based algorithms show limited effectiveness in leveraging social structures to enhance performance. Overall, AdaSociety serves as a valuable research platform for exploring intelligence in diverse physical and social settings. The code is available at https://github.com/bigai-ai/AdaSociety.
comment: Accepted at NeurIPS D&B 2024
♻ ☆ Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space NeurIPS 2024
Modern generative models demonstrate impressive capabilities, likely stemming from an ability to identify and manipulate abstract concepts underlying their training data. However, fundamental questions remain: what determines the concepts a model learns, the order in which it learns them, and its ability to manipulate those concepts? To address these questions, we propose analyzing a model's learning dynamics via a framework we call the concept space, where each axis represents an independent concept underlying the data generating process. By characterizing learning dynamics in this space, we identify how the speed at which a concept is learned, and hence the order of concept learning, is controlled by properties of the data we term concept signal. Further, we observe moments of sudden turns in the direction of a model's learning dynamics in concept space. Surprisingly, these points precisely correspond to the emergence of hidden capabilities, i.e., where latent interventions show the model possesses the capability to manipulate a concept, but these capabilities cannot yet be elicited via naive input prompting. While our results focus on synthetically defined toy datasets, we hypothesize a general claim on emergence of hidden capabilities may hold: generative models possess latent capabilities that emerge suddenly and consistently during training, though a model might not exhibit these capabilities under naive input prompting.
comment: NeurIPS 2024 (Spotlight)
♻ ☆ Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models
We present Deep Compression Autoencoder (DC-AE), a new family of autoencoder models for accelerating high-resolution diffusion models. Existing autoencoder models have demonstrated impressive results at a moderate spatial compression ratio (e.g., 8x), but fail to maintain satisfactory reconstruction accuracy for high spatial compression ratios (e.g., 64x). We address this challenge by introducing two key techniques: (1) Residual Autoencoding, where we design our models to learn residuals based on the space-to-channel transformed features to alleviate the optimization difficulty of high spatial-compression autoencoders; (2) Decoupled High-Resolution Adaptation, an efficient decoupled three-phases training strategy for mitigating the generalization penalty of high spatial-compression autoencoders. With these designs, we improve the autoencoder's spatial compression ratio up to 128 while maintaining the reconstruction quality. Applying our DC-AE to latent diffusion models, we achieve significant speedup without accuracy drop. For example, on ImageNet 512x512, our DC-AE provides 19.1x inference speedup and 17.9x training speedup on H100 GPU for UViT-H while achieving a better FID, compared with the widely used SD-VAE-f8 autoencoder. Our code is available at https://github.com/mit-han-lab/efficientvit.
comment: Preprint. First two authors contributed equally to this work. Update: add diffusion model scaling results
♻ ☆ EARN Fairness: Explaining, Asking, Reviewing, and Negotiating Artificial Intelligence Fairness Metrics Among Stakeholders
Numerous fairness metrics have been proposed and employed by artificial intelligence (AI) experts to quantitatively measure bias and define fairness in AI models. Recognizing the need to accommodate stakeholders' diverse fairness understandings, efforts are underway to solicit their input. However, conveying AI fairness metrics to stakeholders without AI expertise, capturing their personal preferences, and seeking a collective consensus remain challenging and underexplored. To bridge this gap, we propose a new framework, EARN Fairness, which facilitates collective metric decisions among stakeholders without requiring AI expertise. The framework features an adaptable interactive system and a stakeholder-centered EARN Fairness process to Explain fairness metrics, Ask stakeholders' personal metric preferences, Review metrics collectively, and Negotiate a consensus on metric selection. To gather empirical results, we applied the framework to a credit rating scenario and conducted a user study involving 18 decision subjects without AI knowledge. We identify their personal metric preferences and their acceptable level of unfairness in individual sessions. Subsequently, we uncovered how they reached metric consensus in team sessions. Our work shows that the EARN Fairness framework enables stakeholders to express personal preferences and reach consensus, providing practical guidance for implementing human-centered AI fairness in high-risk contexts. Through this approach, we aim to harmonize fairness expectations of diverse stakeholders, fostering more equitable and inclusive AI fairness.
♻ ☆ AFD: Mitigating Feature Gap for Adversarial Robustness by Feature Disentanglement
Adversarial fine-tuning methods enhance adversarial robustness via fine-tuning the pre-trained model in an adversarial training manner. However, we identify that some specific latent features of adversarial samples are confused by adversarial perturbation and lead to an unexpectedly increasing gap between features in the last hidden layer of natural and adversarial samples. To address this issue, we propose a disentanglement-based approach to explicitly model and further remove the specific latent features. We introduce a feature disentangler to separate out the specific latent features from the features of the adversarial samples, thereby boosting robustness by eliminating the specific latent features. Besides, we align clean features in the pre-trained model with features of adversarial samples in the fine-tuned model, to benefit from the intrinsic features of natural samples. Empirical evaluations on three benchmark datasets demonstrate that our approach surpasses existing adversarial fine-tuning methods and adversarial training baselines.
comment: 7 pages, 5 figures
♻ ☆ Symbolic Regression with a Learned Concept Library NeurIPS
We present a novel method for symbolic regression (SR), the task of searching for compact programmatic hypotheses that best explain a dataset. The problem is commonly solved using genetic algorithms; we show that we can enhance such methods by inducing a library of abstract textual concepts. Our algorithm, called LaSR, uses zero-shot queries to a large language model (LLM) to discover and evolve concepts occurring in known high-performing hypotheses. We discover new hypotheses using a mix of standard evolutionary steps and LLM-guided steps (obtained through zero-shot LLM queries) conditioned on discovered concepts. Once discovered, hypotheses are used in a new round of concept abstraction and evolution. We validate LaSR on the Feynman equations, a popular SR benchmark, as well as a set of synthetic tasks. On these benchmarks, LaSR substantially outperforms a variety of state-of-the-art SR approaches based on deep learning and evolutionary algorithms. Moreover, we show that LaSR can be used to discover a novel and powerful scaling law for LLMs.
comment: NeurIPS version; 10 pages; no checklist; added more experiment details
♻ ☆ AFFSegNet: Adaptive Feature Fusion Segmentation Network for Microtumors and Multi-Organ Segmentation
Medical image segmentation, a crucial task in computer vision, facilitates the automated delineation of anatomical structures and pathologies, supporting clinicians in diagnosis, treatment planning, and disease monitoring. Notably, transformers employing shifted window-based self-attention have demonstrated exceptional performance. However, their reliance on local window attention limits the fusion of local and global contextual information, crucial for segmenting microtumors and miniature organs. To address this limitation, we propose the Adaptive Semantic Segmentation Network (ASSNet), a transformer architecture that effectively integrates local and global features for precise medical image segmentation. ASSNet comprises a transformer-based U-shaped encoder-decoder network. The encoder utilizes shifted window self-attention across five resolutions to extract multi-scale features, which are then propagated to the decoder through skip connections. We introduce an augmented multi-layer perceptron within the encoder to explicitly model long-range dependencies during feature extraction. Recognizing the constraints of conventional symmetrical encoder-decoder designs, we propose an Adaptive Feature Fusion (AFF) decoder to complement our encoder. This decoder incorporates three key components: the Long Range Dependencies (LRD) block, the Multi-Scale Feature Fusion (MFF) block, and the Adaptive Semantic Center (ASC) block. These components synergistically facilitate the effective fusion of multi-scale features extracted by the decoder while capturing long-range dependencies and refining object boundaries. Comprehensive experiments on diverse medical image segmentation tasks, including multi-organ, liver tumor, and bladder tumor segmentation, demonstrate that ASSNet achieves state-of-the-art results. Code and models are available at: \url{https://github.com/lzeeorno/ASSNet}.
comment: 8 pages, 4 figures, 3 tables
♻ ☆ Unlocking Feature Visualization for Deeper Networks with MAgnitude Constrained Optimization
Feature visualization has gained substantial popularity, particularly after the influential work by Olah et al. in 2017, which established it as a crucial tool for explainability. However, its widespread adoption has been limited due to a reliance on tricks to generate interpretable images, and corresponding challenges in scaling it to deeper neural networks. Here, we describe MACO, a simple approach to address these shortcomings. The main idea is to generate images by optimizing the phase spectrum while keeping the magnitude constant to ensure that generated explanations lie in the space of natural images. Our approach yields significantly better results (both qualitatively and quantitatively) and unlocks efficient and interpretable feature visualizations for large state-of-the-art neural networks. We also show that our approach exhibits an attribution mechanism allowing us to augment feature visualizations with spatial importance. We validate our method on a novel benchmark for comparing feature visualization methods, and release its visualizations for all classes of the ImageNet dataset on https://serre-lab.github.io/Lens/. Overall, our approach unlocks, for the first time, feature visualizations for large, state-of-the-art deep neural networks without resorting to any parametric prior image model.
♻ ☆ SpaFL: Communication-Efficient Federated Learning with Sparse Models and Low computational Overhead NeurIPS 2024
The large communication and computation overhead of federated learning (FL) is one of the main challenges facing its practical deployment over resource-constrained clients and systems. In this work, SpaFL: a communication-efficient FL framework is proposed to optimize sparse model structures with low computational overhead. In SpaFL, a trainable threshold is defined for each filter/neuron to prune its all connected parameters, thereby leading to structured sparsity. To optimize the pruning process itself, only thresholds are communicated between a server and clients instead of parameters, thereby learning how to prune. Further, global thresholds are used to update model parameters by extracting aggregated parameter importance. The generalization bound of SpaFL is also derived, thereby proving key insights on the relation between sparsity and performance. Experimental results show that SpaFL improves accuracy while requiring much less communication and computing resources compared to sparse baselines. The code is available at https://github.com/news-vt/SpaFL_NeruIPS_2024
comment: Published in NeurIPS 2024
♻ ☆ Electrocardiogram (ECG) Based Cardiac Arrhythmia Detection and Classification using Machine Learning Algorithms
The rapid advancements in Artificial Intelligence, specifically Machine Learning (ML) and Deep Learning (DL), have opened new prospects in medical sciences for improved diagnosis, prognosis, and treatment of severe health conditions. This paper focuses on the development of an ML model with high predictive accuracy to classify arrhythmic electrocardiogram (ECG) signals. The ECG signals datasets utilized in this study were sourced from the PhysioNet and MIT-BIH databases. The research commenced with binary classification, where an optimized Bidirectional Long Short-Term Memory (Bi-LSTM) model yielded excellent results in differentiating normal and atrial fibrillation signals. A pivotal aspect of this research was a survey among medical professionals, which not only validated the practicality of AI-based ECG classifiers but also identified areas for improvement, including accuracy and the inclusion of more arrhythmia types. These insights drove the development of an advanced Convolutional Neural Network (CNN) system capable of classifying five different types of ECG signals with better accuracy and precision. The CNN model's robust performance was ensured through rigorous stratified 5-fold cross validation. A web portal was also developed to demonstrate real-world utility, offering access to the trained model for real-time classification. This study highlights the potential applications of such models in remote health monitoring, predictive healthcare, assistive diagnostic tools, and simulated environments for educational training and interdisciplinary collaboration between data scientists and medical personnel.
♻ ☆ BudgetFusion: Perceptually-Guided Adaptive Diffusion Models
Diffusion models have shown unprecedented success in the task of text-to-image generation. While these models are capable of generating high-quality and realistic images, the complexity of sequential denoising has raised societal concerns regarding high computational demands and energy consumption. In response, various efforts have been made to improve inference efficiency. However, most of the existing efforts have taken a fixed approach with neural network simplification or text prompt optimization. Are the quality improvements from all denoising computations equally perceivable to humans? We observed that images from different text prompts may require different computational efforts given the desired content. The observation motivates us to present BudgetFusion, a novel model that suggests the most perceptually efficient number of diffusion steps before a diffusion model starts to generate an image. This is achieved by predicting multi-level perceptual metrics relative to diffusion steps. With the popular Stable Diffusion as an example, we conduct both numerical analyses and user studies. Our experiments show that BudgetFusion saves up to five seconds per prompt without compromising perceptual similarity. We hope this work can initiate efforts toward answering a core question: how much do humans perceptually gain from images created by a generative model, per watt of energy?
♻ ☆ GenderAlign: An Alignment Dataset for Mitigating Gender Bias in Large Language Models
Large Language Models (LLMs) are prone to generating content that exhibits gender biases, raising significant ethical concerns. Alignment, the process of fine-tuning LLMs to better align with desired behaviors, is recognized as an effective approach to mitigate gender biases. Although proprietary LLMs have made significant strides in mitigating gender bias, their alignment datasets are not publicly available. The commonly used and publicly available alignment dataset, HH-RLHF, still exhibits gender bias to some extent. There is a lack of publicly available alignment datasets specifically designed to address gender bias. Hence, we developed a new dataset named GenderAlign, aiming at mitigating a comprehensive set of gender biases in LLMs. This dataset comprises 8k single-turn dialogues, each paired with a "chosen" and a "rejected" response. Compared to the "rejected" responses, the "chosen" responses demonstrate lower levels of gender bias and higher quality. Furthermore, we categorized the gender biases in the "rejected" responses of GenderAlign into 4 principal categories. The experimental results show the effectiveness of GenderAlign in reducing gender bias in LLMs.
♻ ☆ Guiding Vision-Language Model Selection for Visual Question-Answering Across Tasks, Domains, and Knowledge Types
Visual Question-Answering (VQA) has become key to user experience, particularly after improved generalization capabilities of Vision-Language Models (VLMs). But evaluating VLMs for an application requirement using a standardized framework in practical settings is still challenging. This paper aims to solve that using an end-to-end framework. We present VQA360 - a novel dataset derived from established VQA benchmarks, annotated with task types, application domains, and knowledge types, for a comprehensive evaluation. We also introduce GoEval, a multimodal evaluation metric developed using GPT-4o, achieving a correlation factor of 56.71% with human judgments. Our experiments with state-of-the-art VLMs reveal that no single model excels universally, thus, making a right choice a key design decision. Proprietary models such as Gemini-1.5-Pro and GPT-4o-mini generally outperform others, but open-source models like InternVL-2-8B and CogVLM-2-Llama-3-19B also demonstrate competitive strengths, while providing additional advantages. Our framework can also be extended to other tasks.
comment: 8 pages + references + 6 pages of Appendix
♻ ☆ [MASK] is All You Need
In generative models, two paradigms have gained attraction in various applications: next-set prediction-based Masked Generative Models and next-noise prediction-based Non-Autoregressive Models, e.g., Diffusion Models. In this work, we propose using discrete-state models to connect them and explore their scalability in the vision domain. First, we conduct a step-by-step analysis in a unified design space across two types of models including timestep-independence, noise schedule, temperature, guidance strength, etc in a scalable manner. Second, we re-cast typical discriminative tasks, e.g., image segmentation, as an unmasking process from [MASK] tokens on a discrete-state model. This enables us to perform various sampling processes, including flexible conditional sampling by only training once to model the joint distribution. All aforementioned explorations lead to our framework named Discrete Interpolants, which enables us to achieve state-of-the-art or competitive performance compared to previous discrete-state based methods in various benchmarks, like ImageNet256, MS COCO, and video dataset FaceForensics. In summary, by leveraging [MASK] in discrete-state models, we can bridge Masked Generative and Non-autoregressive Diffusion models, as well as generative and discriminative tasks.
comment: Technical Report (WIP), Project Page(code, model, dataset): https://compvis.github.io/mask/
♻ ☆ Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Change KDD
Counterfactual explanations (CFEs) guide users on how to adjust inputs to machine learning models to achieve desired outputs. While existing research primarily addresses static scenarios, real-world applications often involve data or model changes, potentially invalidating previously generated CFEs and rendering user-induced input changes ineffective. Current methods addressing this issue often support only specific models or change types, require extensive hyperparameter tuning, or fail to provide probabilistic guarantees on CFE robustness to model changes. This paper proposes a novel approach for generating CFEs that provides probabilistic guarantees for any model and change type, while offering interpretable and easy-to-select hyperparameters. We establish a theoretical framework for probabilistically defining robustness to model change and demonstrate how our BetaRCE method directly stems from it. BetaRCE is a post-hoc method applied alongside a chosen base CFE generation method to enhance the quality of the explanation beyond robustness. It facilitates a transition from the base explanation to a more robust one with user-adjusted probability bounds. Through experimental comparisons with baselines, we show that BetaRCE yields robust, most plausible, and closest to baseline counterfactual explanations.
comment: Accepted at 31st SIGKDD Conference on Knowledge Discovery and Data Mining 2025
♻ ☆ AutoDetect: Towards a Unified Framework for Automated Weakness Detection in Large Language Models EMNLP 2024
Although Large Language Models (LLMs) are becoming increasingly powerful, they still exhibit significant but subtle weaknesses, such as mistakes in instruction-following or coding tasks. As these unexpected errors could lead to severe consequences in practical deployments, it is crucial to investigate the limitations within LLMs systematically. Traditional benchmarking approaches cannot thoroughly pinpoint specific model deficiencies, while manual inspections are costly and not scalable. In this paper, we introduce a unified framework, AutoDetect, to automatically expose weaknesses in LLMs across various tasks. Inspired by the educational assessment process that measures students' learning outcomes, AutoDetect consists of three LLM-powered agents: Examiner, Questioner, and Assessor. The collaboration among these three agents is designed to realize comprehensive and in-depth weakness identification. Our framework demonstrates significant success in uncovering flaws, with an identification success rate exceeding 30% in prominent models such as ChatGPT and Claude. More importantly, these identified weaknesses can guide specific model improvements, proving more effective than untargeted data augmentation methods like Self-Instruct. Our approach has led to substantial enhancements in popular LLMs, including the Llama series and Mistral-7b, boosting their performance by over 10% across several benchmarks. Code and data are publicly available at https://github.com/thu-coai/AutoDetect.
comment: EMNLP 2024 findings
♻ ☆ ContRail: A Framework for Realistic Railway Image Synthesis using ControlNet
Deep Learning became an ubiquitous paradigm due to its extraordinary effectiveness and applicability in numerous domains. However, the approach suffers from the high demand of data required to achieve the potential of this type of model. An ever-increasing sub-field of Artificial Intelligence, Image Synthesis, aims to address this limitation through the design of intelligent models capable of creating original and realistic images, endeavour which could drastically reduce the need for real data. The Stable Diffusion generation paradigm recently propelled state-of-the-art approaches to exceed all previous benchmarks. In this work, we propose the ContRail framework based on the novel Stable Diffusion model ControlNet, which we empower through a multi-modal conditioning method. We experiment with the task of synthetic railway image generation, where we improve the performance in rail-specific tasks, such as rail semantic segmentation by enriching the dataset with realistic synthetic images.
comment: 9 pages, 5 figures, 2 tables
♻ ☆ Separation Power of Equivariant Neural Networks
The separation power of a machine learning model refers to its ability to distinguish between different inputs and is often used as a proxy for its expressivity. Indeed, knowing the separation power of a family of models is a necessary condition to obtain fine-grained universality results. In this paper, we analyze the separation power of equivariant neural networks, such as convolutional and permutation-invariant networks. We first present a complete characterization of inputs indistinguishable by models derived by a given architecture. From this results, we derive how separability is influenced by hyperparameters and architectural choices-such as activation functions, depth, hidden layer width, and representation types. Notably, all non-polynomial activations, including ReLU and sigmoid, are equivalent in expressivity and reach maximum separation power. Depth improves separation power up to a threshold, after which further increases have no effect. Adding invariant features to hidden representations does not impact separation power. Finally, block decomposition of hidden representations affects separability, with minimal components forming a hierarchy in separation power that provides a straightforward method for comparing the separation power of models.
comment: 10 pages of main text, 1 figure
♻ ☆ Enhancing Vision-Language Model Pre-training with Image-text Pair Pruning Based on Word Frequency
We propose Word-Frequency-based Image-Text Pair Pruning (WFPP), a novel data pruning method that improves the efficiency of VLMs. Unlike MetaCLIP, our method does not need metadata for pruning, but selects text-image pairs to prune based on the content of the text. Specifically, WFPP prunes text-image pairs containing high-frequency words across the entire training dataset. The effect of WFPP is to reduce the dominance of frequent words. The result a better balanced word-frequency distribution in the dataset, which is known to improve the training of word embedding models. After pre-training on the pruned subset, we fine-tuned the model on the entire dataset for one additional epoch to achieve better performance. Our experiments demonstrate that applying WFPP when training a CLIP model improves performance on a wide range of downstream tasks. WFPP also provides the advantage of speeding up pre-training by using fewer samples. Additionally, we analyze the training data before and after pruning to visualize how WFPP changes the balance of word frequencies. We hope our work encourages researchers to consider the distribution of words in the training data when pre-training VLMs, not limited to CLIP.
♻ ☆ S+t-SNE -- Bringing Dimensionality Reduction to Data Streams
We present S+t-SNE, an adaptation of the t-SNE algorithm designed to handle infinite data streams. The core idea behind S+t-SNE is to update the t-SNE embedding incrementally as new data arrives, ensuring scalability and adaptability to handle streaming scenarios. By selecting the most important points at each step, the algorithm ensures scalability while keeping informative visualisations. By employing a blind method for drift management, the algorithm adjusts the embedding space, which facilitates the visualisation of evolving data dynamics. Our experimental evaluations demonstrate the effectiveness and efficiency of S+t-SNE, whilst highlighting its ability to capture patterns in a streaming scenario. We hope our approach offers researchers and practitioners a real-time tool for understanding and interpreting high-dimensional data.
comment: This preprint has undergone peer review but does not have any post-submission improvements or corrections. Full version after peer-review and post-acceptance improvements was presented at IDA2024 (https://ida2024.org/)
♻ ☆ Unveiling Environmental Sensitivity of Individual Gains in Influence Maximization
Influence Maximization (IM) is to identify the seed set to maximize information dissemination in a network. Elegant IM algorithms could naturally extend to cases where each node is equipped with a specific weight, reflecting individual gains to measure the node's importance. Prevailing literature typically assumes such individual gains remain constant throughout the cascade process and are solvable through explicit formulas based on the node's characteristics and network topology. However, this assumption is not always feasible for two reasons: 1)Unobservability: The individual gains of each node are primarily evaluated by the difference between the outputs in the activated and non-activated states. In practice, we can only observe one of these states, with the other remaining unobservable post-propagation. 2)Environmental sensitivity: In addition to the node's inherent properties, individual gains are also sensitive to the activation status of surrounding nodes, which is dynamic during iteration even when the network topology remains static. To address these challenges, we extend the consideration of IM to a broader scenario with dynamic node individual gains, leveraging causality techniques. In our paper, we introduce a Causal Influence Maximization (CauIM) framework and develop two algorithms, G-CauIM and A-CauIM, where the latter incorporates a novel acceleration technique. Theoretically, we establish the generalized lower bound of influence spread and provide robustness analysis. Empirically, in synthetic and real-world experiments, we demonstrate the effectiveness and robustness of our algorithms.
♻ ☆ pfl-research: simulation framework for accelerating research in Private Federated Learning
Federated learning (FL) is an emerging machine learning (ML) training paradigm where clients own their data and collaborate to train a global model, without revealing any data to the server and other participants. Researchers commonly perform experiments in a simulation environment to quickly iterate on ideas. However, existing open-source tools do not offer the efficiency required to simulate FL on larger and more realistic FL datasets. We introduce pfl-research, a fast, modular, and easy-to-use Python framework for simulating FL. It supports TensorFlow, PyTorch, and non-neural network models, and is tightly integrated with state-of-the-art privacy algorithms. We study the speed of open-source FL frameworks and show that pfl-research is 7-72$\times$ faster than alternative open-source frameworks on common cross-device setups. Such speedup will significantly boost the productivity of the FL research community and enable testing hypotheses on realistic FL datasets that were previously too resource intensive. We release a suite of benchmarks that evaluates an algorithm's overall performance on a diverse set of realistic scenarios. The code is available on GitHub at https://github.com/apple/pfl-research.
♻ ☆ Hypergrah-Enhanced Dual Convolutional Network for Bundle Recommendation
Bundle recommendations strive to offer users a set of items as a package named bundle, enhancing convenience and contributing to the seller's revenue. While previous approaches have demonstrated notable performance, we argue that they may compromise the ternary relationship among users, items, and bundles. This compromise can result in information loss, ultimately impacting the overall model performance. To address this gap, we develop a unified model for bundle recommendation, termed hypergraph-enhanced dual convolutional neural network (HED). Our approach is characterized by two key aspects. Firstly, we construct a complete hypergraph to capture interaction dynamics among users, items, and bundles. Secondly, we incorporate U-B interaction information to enhance the information representation derived from users and bundle embedding vectors. Extensive experimental results on the Youshu and Netease datasets have demonstrated that HED surpasses state-of-the-art baselines, proving its effectiveness. In addition, various ablation studies and sensitivity analyses revealed the working mechanism and proved our effectiveness. Codes and datasets are available at https://github.com/AAI-Lab/HED
♻ ☆ Active Inference Tree Search in Large POMDPs
The ability to plan ahead efficiently is key for both living organisms and artificial systems. Model-based planning and prospection are widely studied in cognitive neuroscience and artificial intelligence (AI), but from different perspectives--and with different desiderata in mind (biological realism versus scalability) that are difficult to reconcile. Here, we introduce a novel method to plan in POMDPs--Active Inference Tree Search (AcT)--that combines the normative character and biological realism of a leading planning theory in neuroscience (Active Inference) and the scalability of tree search methods in AI. This unification enhances both approaches. On the one hand, tree searches enable the biologically grounded, first principle method of active inference to be applied to large-scale problems. On the other hand, active inference provides a principled solution to the exploration-exploitation dilemma, which is often addressed heuristically in tree search methods. Our simulations show that AcT successfully navigates binary trees that are challenging for sampling-based methods, problems that require adaptive exploration, and the large POMDP problem 'RockSample'--in which AcT reproduces state-of-the-art POMDP solutions. Furthermore, we illustrate how AcT can be used to simulate neurophysiological responses (e.g., in the hippocampus and prefrontal cortex) of humans and other animals that solve large planning problems. These numerical analyses show that Active Tree Search is a principled realisation of neuroscientific and AI planning theories, which offer both biological realism and scalability.
comment: 47 pages, 9 figures, 1 Appendix of two sections with pseudocodes and one encoding example, submitted preprint
♻ ☆ Robust Markov Decision Processes: A Place Where AI and Formal Methods Meet
Markov decision processes (MDPs) are a standard model for sequential decision-making problems and are widely used across many scientific areas, including formal methods and artificial intelligence (AI). MDPs do, however, come with the restrictive assumption that the transition probabilities need to be precisely known. Robust MDPs (RMDPs) overcome this assumption by instead defining the transition probabilities to belong to some uncertainty set. We present a gentle survey on RMDPs, providing a tutorial covering their fundamentals. In particular, we discuss RMDP semantics and how to solve them by extending standard MDP methods such as value iteration and policy iteration. We also discuss how RMDPs relate to other models and how they are used in several contexts, including reinforcement learning and abstraction techniques. We conclude with some challenges for future work on RMDPs.
♻ ☆ Software Model Evolution with Large Language Models: Experiments on Simulated, Public, and Industrial Datasets
Modeling structure and behavior of software systems plays a crucial role in the industrial practice of software engineering. As with other software engineering artifacts, software models are subject to evolution. Supporting modelers in evolving software models with recommendations for model completions is still an open problem, though. In this paper, we explore the potential of large language models for this task. In particular, we propose an approach, RAMC, leveraging large language models, model histories, and retrieval-augmented generation for model completion. Through experiments on three datasets, including an industrial application, one public open-source community dataset, and one controlled collection of simulated model repositories, we evaluate the potential of large language models for model completion with RAMC. We found that large language models are indeed a promising technology for supporting software model evolution (62.30% semantically correct completions on real-world industrial data and up to 86.19% type-correct completions). The general inference capabilities of large language models are particularly useful when dealing with concepts for which there are few, noisy, or no examples at all.
♻ ☆ A Comprehensive Survey on Kolmogorov Arnold Networks (KAN)
Through this comprehensive survey of Kolmogorov-Arnold Networks(KAN), we have gained a thorough understanding of its theoretical foundation, architectural design, application scenarios, and current research progress. KAN, with its unique architecture and flexible activation functions, excels in handling complex data patterns and nonlinear relationships, demonstrating wide-ranging application potential. While challenges remain, KAN is poised to pave the way for innovative solutions in various fields, potentially revolutionizing how we approach complex computational problems.
♻ ☆ RST-LoRA: A Discourse-Aware Low-Rank Adaptation for Long Document Abstractive Summarization NAACL 2024
For long document summarization, discourse structure is important to discern the key content of the text and the differences in importance level between sentences. Unfortunately, the integration of rhetorical structure theory (RST) into parameter-efficient fine-tuning strategies for long document summarization remains unexplored. Therefore, this paper introduces RST-LoRA and proposes four RST-aware variants to explicitly incorporate RST into the LoRA model. Our empirical evaluation demonstrates that incorporating the type and uncertainty of rhetorical relations can complementarily enhance the performance of LoRA in summarization tasks. Furthermore, the best-performing variant we introduced outperforms the vanilla LoRA and full-parameter fine-tuning models, as confirmed by multiple automatic and human evaluations, and even surpasses previous state-of-the-art methods.
comment: NAACL 2024 Main & Long Conference Paper (Oral Presentation)
♻ ☆ SciNews: From Scholarly Complexities to Public Narratives -- A Dataset for Scientific News Report Generation LREC
Scientific news reports serve as a bridge, adeptly translating complex research articles into reports that resonate with the broader public. The automated generation of such narratives enhances the accessibility of scholarly insights. In this paper, we present a new corpus to facilitate this paradigm development. Our corpus comprises a parallel compilation of academic publications and their corresponding scientific news reports across nine disciplines. To demonstrate the utility and reliability of our dataset, we conduct an extensive analysis, highlighting the divergences in readability and brevity between scientific news narratives and academic manuscripts. We benchmark our dataset employing state-of-the-art text generation models. The evaluation process involves both automatic and human evaluation, which lays the groundwork for future explorations into the automated generation of scientific news reports. The dataset and code related to this work are available at https://dongqi.me/projects/SciNews.
comment: LREC-COLING 2024 Main Conference Paper
♻ ☆ ChatGPT vs Human-authored Text: Insights into Controllable Text Summarization and Sentence Style Transfer ACL
Large-scale language models, like ChatGPT, have garnered significant media attention and stunned the public with their remarkable capacity for generating coherent text from short natural language prompts. In this paper, we aim to conduct a systematic inspection of ChatGPT's performance in two controllable generation tasks, with respect to ChatGPT's ability to adapt its output to different target audiences (expert vs. layman) and writing styles (formal vs. informal). Additionally, we evaluate the faithfulness of the generated text, and compare the model's performance with human-authored texts. Our findings indicate that the stylistic variations produced by humans are considerably larger than those demonstrated by ChatGPT, and the generated texts diverge from human samples in several characteristics, such as the distribution of word types. Moreover, we observe that ChatGPT sometimes incorporates factual errors or hallucinations when adapting the text to suit a specific style.
comment: ACL-SRW 2023
♻ ☆ Incorporating Distributions of Discourse Structure for Long Document Abstractive Summarization ACL 2023
For text summarization, the role of discourse structure is pivotal in discerning the core content of a text. Regrettably, prior studies on incorporating Rhetorical Structure Theory (RST) into transformer-based summarization models only consider the nuclearity annotation, thereby overlooking the variety of discourse relation types. This paper introduces the 'RSTformer', a novel summarization model that comprehensively incorporates both the types and uncertainty of rhetorical relations. Our RST-attention mechanism, rooted in document-level rhetorical structure, is an extension of the recently devised Longformer framework. Through rigorous evaluation, the model proposed herein exhibits significant superiority over state-of-the-art models, as evidenced by its notable performance on several automatic metrics and human evaluation.
comment: ACL 2023 (Main conference)
♻ ☆ Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision NeurIPS 2024
Current AI alignment methodologies rely on human-provided demonstrations or judgments, and the learned capabilities of AI systems would be upper-bounded by human capabilities as a result. This raises a challenging research question: How can we keep improving the systems when their capabilities have surpassed the levels of humans? This paper answers this question in the context of tackling hard reasoning tasks (e.g., level 4-5 MATH problems) via learning from human annotations on easier tasks (e.g., level 1-3 MATH problems), which we term as easy-to-hard generalization. Our key insight is that an evaluator (reward model) trained on supervisions for easier tasks can be effectively used for scoring candidate solutions of harder tasks and hence facilitating easy-to-hard generalization over different levels of tasks. Based on this insight, we propose a novel approach to scalable alignment, which firstly trains the (process-supervised) reward models on easy problems (e.g., level 1-3), and then uses them to evaluate the performance of policy models on hard problems. We show that such easy-to-hard generalization from evaluators can enable easy-to-hard generalizations in generators either through re-ranking or reinforcement learning (RL). Notably, our process-supervised 7b RL model and 34b model (reranking@1024) achieves an accuracy of 34.0% and 52.5% on MATH500, respectively, despite only using human supervision on easy problems. Our approach suggests a promising path toward AI systems that advance beyond the frontier of human supervision.
comment: Accepted at NeurIPS 2024
♻ ☆ Detecting and Corrupting Convolution-based Unlearnable Examples AAAI 2025
Convolution-based unlearnable examples (UEs) employ class-wise multiplicative convolutional noise to training samples, severely compromising model performance. This fire-new type of UEs have successfully countered all defense mechanisms against UEs. The failure of such defenses can be attributed to the absence of norm constraints on convolutional noise, leading to severe blurring of image features. To address this, we first design an Edge Pixel-based Detector (EPD) to identify convolution-based UEs. Upon detection of them, we propose the first defense scheme against convolution-based UEs, COrrupting these samples via random matrix multiplication by employing bilinear INterpolation (COIN) such that disrupting the distribution of class-wise multiplicative noise. To evaluate the generalization of our proposed COIN, we newly design two convolution-based UEs called VUDA and HUDA to expand the scope of convolution-based UEs. Extensive experiments demonstrate the effectiveness of detection scheme EPD and that our defense COIN outperforms 11 state-of-the-art (SOTA) defenses, achieving a significant improvement on the CIFAR and ImageNet datasets.
comment: AAAI 2025
♻ ☆ ProcessBench: Identifying Process Errors in Mathematical Reasoning
As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.
♻ ☆ DSTC: Direct Preference Learning with Only Self-Generated Tests and Code to Improve Code LMs
Direct preference learning offers a promising and computation-efficient beyond supervised fine-tuning (SFT) for improving code generation in coding large language models (LMs). However, the scarcity of reliable preference data is a bottleneck for the performance of direct preference learning to improve the coding accuracy of code LMs. In this paper, we introduce \underline{\textbf{D}}irect Preference Learning with Only \underline{\textbf{S}}elf-Generated \underline{\textbf{T}}ests and \underline{\textbf{C}}ode (DSTC), a framework that leverages only self-generated code snippets and tests to construct reliable preference pairs such that direct preference learning can improve LM coding accuracy without external annotations. DSTC combines a minimax selection process and test-code concatenation to improve preference pair quality, reducing the influence of incorrect self-generated tests and enhancing model performance without the need for costly reward models. When applied with direct preference learning methods such as Direct Preference Optimization (DPO) and Kahneman-Tversky Optimization (KTO), DSTC yields stable improvements in coding accuracy (pass@1 score) across diverse coding benchmarks, including HumanEval, MBPP, and BigCodeBench, demonstrating both its effectiveness and scalability for models of various sizes. This approach autonomously enhances code generation accuracy across LLMs of varying sizes, reducing reliance on expensive annotated coding datasets.
♻ ☆ LEARN: Knowledge Adaptation from Large Language Model to Recommendation for Practical Industrial Application AAAI 2025
Contemporary recommendation systems predominantly rely on ID embedding to capture latent associations among users and items. However, this approach overlooks the wealth of semantic information embedded within textual descriptions of items, leading to suboptimal performance and poor generalizations. Leveraging the capability of large language models to comprehend and reason about textual content presents a promising avenue for advancing recommendation systems. To achieve this, we propose an Llm-driven knowlEdge Adaptive RecommeNdation (LEARN) framework that synergizes open-world knowledge with collaborative knowledge. We address computational complexity concerns by utilizing pretrained LLMs as item encoders and freezing LLM parameters to avoid catastrophic forgetting and preserve open-world knowledge. To bridge the gap between the open-world and collaborative domains, we design a twin-tower structure supervised by the recommendation task and tailored for practical industrial application. Through experiments on the real large-scale industrial dataset and online A/B tests, we demonstrate the efficacy of our approach in industry application. We also achieve state-of-the-art performance on six Amazon Review datasets to verify the superiority of our method.
comment: Accepted by AAAI 2025
♻ ☆ Track4Gen: Teaching Video Diffusion Models to Track Points Improves Video Generation
While recent foundational video generators produce visually rich output, they still struggle with appearance drift, where objects gradually degrade or change inconsistently across frames, breaking visual coherence. We hypothesize that this is because there is no explicit supervision in terms of spatial tracking at the feature level. We propose Track4Gen, a spatially aware video generator that combines video diffusion loss with point tracking across frames, providing enhanced spatial supervision on the diffusion features. Track4Gen merges the video generation and point tracking tasks into a single network by making minimal changes to existing video generation architectures. Using Stable Video Diffusion as a backbone, Track4Gen demonstrates that it is possible to unify video generation and point tracking, which are typically handled as separate tasks. Our extensive evaluations show that Track4Gen effectively reduces appearance drift, resulting in temporally stable and visually coherent video generation. Project page: hyeonho99.github.io/track4gen
comment: Project page: hyeonho99.github.io/track4gen
♻ ☆ HAAT: Hybrid Attention Aggregation Transformer for Image Super-Resolution
In the research area of image super-resolution, Swin-transformer-based models are favored for their global spatial modeling and shifting window attention mechanism. However, existing methods often limit self-attention to non overlapping windows to cut costs and ignore the useful information that exists across channels. To address this issue, this paper introduces a novel model, the Hybrid Attention Aggregation Transformer (HAAT), designed to better leverage feature information. HAAT is constructed by integrating Swin-Dense-Residual-Connected Blocks (SDRCB) with Hybrid Grid Attention Blocks (HGAB). SDRCB expands the receptive field while maintaining a streamlined architecture, resulting in enhanced performance. HGAB incorporates channel attention, sparse attention, and window attention to improve nonlocal feature fusion and achieve more visually compelling results. Experimental evaluations demonstrate that HAAT surpasses state-of-the-art methods on benchmark datasets. Keywords: Image super-resolution, Computer vision, Attention mechanism, Transformer
comment: 6 pages, 2 figures, 1 table
♻ ☆ Policy Filtration in RLHF to Fine-Tune LLM for Code Generation
Reinforcement learning from human feedback (RLHF) is one of the key techniques that helps large language models (LLMs) to follow instructions and provide helpful and harmless responses. While direct policy optimization methods exist, state-of-the-art LLMs adopt RL-based methods (usually PPO) in RLHF to train the policy to generate good responses guided by a reward model learned from preference data. The main challenge of these methods is the inaccuracy of the intermediate reward model, especially in code generation tasks that require long and complex reasoning to score a response. We find that the reliability of the reward model varies across responses assigned with different rewards. This motivates us to filter the samples whose rewards may be unreliable to improve signal-to-noise ratio during policy learning, resulting in Policy Filtration for Proximal Policy Optimization (PF-PPO). To choose a proper policy filtration strategy for a given reward model, the coefficient of determination ($R^2$) between rewards and actual scores on filtered samples serves as a good metrics and helps us find several promising strategies. We provide extensive experiments to validate the effectiveness of PF-PPO in code generation tasks, and find that some variants of PF-PPO are highly effective and achieve new state-of-the-art performance across 7-billion-parameter models on HumanEval, MBPP, and a new and more challenging LeetCode Contest benchmark.
♻ ☆ Distribution-Level Feature Distancing for Machine Unlearning: Towards a Better Trade-off Between Model Utility and Forgetting AAAI 2025
With the explosive growth of deep learning applications and increasing privacy concerns, the right to be forgotten has become a critical requirement in various AI industries. For example, given a facial recognition system, some individuals may wish to remove their personal data that might have been used in the training phase. Unfortunately, deep neural networks sometimes unexpectedly leak personal identities, making this removal challenging. While recent machine unlearning algorithms aim to enable models to forget such data, we observe an unintended utility drop, termed correlation collapse, where these algorithms inadvertently weaken the essential correlations between image features and true labels during the forgetting process. To address this challenge, we propose Distribution-Level Feature Distancing (DLFD), a novel method that efficiently forgets instances while preserving task-relevant feature correlations. Our method synthesizes data samples by optimizing the feature distribution to be distinctly different from that of forget samples, achieving effective results within a single training epoch. Through extensive experiments on facial recognition datasets, we demonstrate that our approach significantly outperforms state-of-the-art machine unlearning methods in both forgetting performance and model utility preservation.
comment: 10 pages, 6 figures, AAAI 2025 camera ready version
♻ ☆ Trusted Unified Feature-Neighborhood Dynamics for Multi-View Classification AAAI 2025
Multi-view classification (MVC) faces inherent challenges due to domain gaps and inconsistencies across different views, often resulting in uncertainties during the fusion process. While Evidential Deep Learning (EDL) has been effective in addressing view uncertainty, existing methods predominantly rely on the Dempster-Shafer combination rule, which is sensitive to conflicting evidence and often neglects the critical role of neighborhood structures within multi-view data. To address these limitations, we propose a Trusted Unified Feature-NEighborhood Dynamics (TUNED) model for robust MVC. This method effectively integrates local and global feature-neighborhood (F-N) structures for robust decision-making. Specifically, we begin by extracting local F-N structures within each view. To further mitigate potential uncertainties and conflicts in multi-view fusion, we employ a selective Markov random field that adaptively manages cross-view neighborhood dependencies. Additionally, we employ a shared parameterized evidence extractor that learns global consensus conditioned on local F-N structures, thereby enhancing the global integration of multi-view features. Experiments on benchmark datasets show that our method improves accuracy and robustness over existing approaches, particularly in scenarios with high uncertainty and conflicting views. The code will be made available at https://github.com/JethroJames/TUNED.
comment: Accepted to AAAI 2025
♻ ☆ From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge
Assessment and evaluation have long been critical challenges in artificial intelligence (AI) and natural language processing (NLP). However, traditional methods, whether matching-based or embedding-based, often fall short of judging subtle attributes and delivering satisfactory results. Recent advancements in Large Language Models (LLMs) inspire the "LLM-as-a-judge" paradigm, where LLMs are leveraged to perform scoring, ranking, or selection across various tasks and applications. This paper provides a comprehensive survey of LLM-based judgment and assessment, offering an in-depth overview to advance this emerging field. We begin by giving detailed definitions from both input and output perspectives. Then we introduce a comprehensive taxonomy to explore LLM-as-a-judge from three dimensions: what to judge, how to judge and where to judge. Finally, we compile benchmarks for evaluating LLM-as-a-judge and highlight key challenges and promising directions, aiming to provide valuable insights and inspire future research in this promising research area. Paper list and more resources about LLM-as-a-judge can be found at \url{https://github.com/llm-as-a-judge/Awesome-LLM-as-a-judge} and \url{https://llm-as-a-judge.github.io}.
comment: v2: add missing citations; 32 pages, 5 figures
♻ ☆ HybridGS: Decoupling Transients and Statics with 2D and 3D Gaussian Splatting
Generating high-quality novel view renderings of 3D Gaussian Splatting (3DGS) in scenes featuring transient objects is challenging. We propose a novel hybrid representation, termed as HybridGS, using 2D Gaussians for transient objects per image and maintaining traditional 3D Gaussians for the whole static scenes. Note that, the 3DGS itself is better suited for modeling static scenes that assume multi-view consistency, but the transient objects appear occasionally and do not adhere to the assumption, thus we model them as planar objects from a single view, represented with 2D Gaussians. Our novel representation decomposes the scene from the perspective of fundamental viewpoint consistency, making it more reasonable. Additionally, we present a novel multi-view regulated supervision method for 3DGS that leverages information from co-visible regions, further enhancing the distinctions between the transients and statics. Then, we propose a straightforward yet effective multi-stage training strategy to ensure robust training and high-quality view synthesis across various settings. Experiments on benchmark datasets show our state-of-the-art performance of novel view synthesis in both indoor and outdoor scenes, even in the presence of distracting elements.
comment: Project page: https://gujiaqivadin.github.io/hybridgs/
♻ ☆ o1-Coder: an o1 Replication for Coding
The technical report introduces O1-CODER, an attempt to replicate OpenAI's o1 model with a focus on coding tasks. It integrates reinforcement learning (RL) and Monte Carlo Tree Search (MCTS) to enhance the model's System-2 thinking capabilities. The framework includes training a Test Case Generator (TCG) for standardized code testing, using MCTS to generate code data with reasoning processes, and iteratively fine-tuning the policy model to initially produce pseudocode and then generate the full code. The report also addresses the opportunities and challenges in deploying o1-like models in real-world applications, suggesting transitioning to the System-2 paradigm and highlighting the imperative for world model construction. Updated model progress and experimental results will be reported in subsequent versions. All source code, curated datasets, as well as the derived models are disclosed at https://github.com/ADaM-BJTU/O1-CODER .
♻ ☆ Leveraging LLM for Automated Ontology Extraction and Knowledge Graph Generation
Extracting relevant and structured knowledge from large, complex technical documents within the Reliability and Maintainability (RAM) domain is labor-intensive and prone to errors. Our work addresses this challenge by presenting OntoKGen, a genuine pipeline for ontology extraction and Knowledge Graph (KG) generation. OntoKGen leverages Large Language Models (LLMs) through an interactive user interface guided by our adaptive iterative Chain of Thought (CoT) algorithm to ensure that the ontology extraction process and, thus, KG generation align with user-specific requirements. Although KG generation follows a clear, structured path based on the confirmed ontology, there is no universally correct ontology as it is inherently based on the user's preferences. OntoKGen recommends an ontology grounded in best practices, minimizing user effort and providing valuable insights that may have been overlooked, all while giving the user complete control over the final ontology. Having generated the KG based on the confirmed ontology, OntoKGen enables seamless integration into schemeless, non-relational databases like Neo4j. This integration allows for flexible storage and retrieval of knowledge from diverse, unstructured sources, facilitating advanced querying, analysis, and decision-making. Moreover, the generated KG serves as a robust foundation for future integration into Retrieval Augmented Generation (RAG) systems, offering enhanced capabilities for developing domain-specific intelligent applications.
♻ ☆ AnomalyControl: Learning Cross-modal Semantic Features for Controllable Anomaly Synthesis
Anomaly synthesis is a crucial approach to augment abnormal data for advancing anomaly inspection. Based on the knowledge from the large-scale pre-training, existing text-to-image anomaly synthesis methods predominantly focus on textual information or coarse-aligned visual features to guide the entire generation process. However, these methods often lack sufficient descriptors to capture the complicated characteristics of realistic anomalies (e.g., the fine-grained visual pattern of anomalies), limiting the realism and generalization of the generation process. To this end, we propose a novel anomaly synthesis framework called AnomalyControl to learn cross-modal semantic features as guidance signals, which could encode the generalized anomaly cues from text-image reference prompts and improve the realism of synthesized abnormal samples. Specifically, AnomalyControl adopts a flexible and non-matching prompt pair (i.e., a text-image reference prompt and a targeted text prompt), where a Cross-modal Semantic Modeling (CSM) module is designed to extract cross-modal semantic features from the textual and visual descriptors. Then, an Anomaly-Semantic Enhanced Attention (ASEA) mechanism is formulated to allow CSM to focus on the specific visual patterns of the anomaly, thus enhancing the realism and contextual relevance of the generated anomaly features. Treating cross-modal semantic features as the prior, a Semantic Guided Adapter (SGA) is designed to encode effective guidance signals for the adequate and controllable synthesis process. Extensive experiments indicate that AnomalyControl can achieve state-of-the-art results in anomaly synthesis compared with existing methods while exhibiting superior performance for downstream tasks.
♻ ☆ S$^{2}$FT: Efficient, Scalable and Generalizable LLM Fine-tuning by Structured Sparsity
Current PEFT methods for LLMs can achieve either high quality, efficient training, or scalable serving, but not all three simultaneously. To address this limitation, we investigate sparse fine-tuning and observe a remarkable improvement in generalization ability. Utilizing this key insight, we propose a family of Structured Sparse Fine-Tuning (S$^{2}$FT) methods for LLMs, which concurrently achieve state-of-the-art fine-tuning performance, training efficiency, and inference scalability. S$^{2}$FT accomplishes this by "selecting sparsely and computing densely". It selects a few heads and channels in the MHA and FFN modules for each Transformer block, respectively. Next, it co-permutes weight matrices on both sides of the coupled structures in LLMs to connect the selected components in each layer into a dense submatrix. Finally, S$^{2}$FT performs in-place gradient updates on all submatrices. Through theoretical analysis and empirical results, our method prevents overfitting and forgetting, delivers SOTA performance on both commonsense and arithmetic reasoning with 4.6% and 1.3% average improvements compared to LoRA, and surpasses full FT by 11.5% when generalizing to various domains after instruction tuning. Using our partial backpropagation algorithm, S$^{2}$FT saves training memory up to 3$\times$ and improves latency by 1.5-2.7$\times$ compared to full FT, while delivering an average 10% improvement over LoRA on both metrics. We further demonstrate that the weight updates in S$^{2}$FT can be decoupled into adapters, enabling effective fusion, fast switch, and efficient parallelism for serving multiple fine-tuned models.
♻ ☆ From Text to Multimodality: Exploring the Evolution and Impact of Large Language Models in Medical Practice
Large Language Models (LLMs) have rapidly evolved from text-based systems to multimodal platforms, significantly impacting various sectors including healthcare. This comprehensive review explores the progression of LLMs to Multimodal Large Language Models (MLLMs) and their growing influence in medical practice. We examine the current landscape of MLLMs in healthcare, analyzing their applications across clinical decision support, medical imaging, patient engagement, and research. The review highlights the unique capabilities of MLLMs in integrating diverse data types, such as text, images, and audio, to provide more comprehensive insights into patient health. We also address the challenges facing MLLM implementation, including data limitations, technical hurdles, and ethical considerations. By identifying key research gaps, this paper aims to guide future investigations in areas such as dataset development, modality alignment methods, and the establishment of ethical guidelines. As MLLMs continue to shape the future of healthcare, understanding their potential and limitations is crucial for their responsible and effective integration into medical practice.
comment: 12 pages, 1 figure
♻ ☆ SAG-ViT: A Scale-Aware, High-Fidelity Patching Approach with Graph Attention for Vision Transformers
Image classification is a computer vision task where a model analyzes an image to categorize it into a specific label. Vision Transformers (ViT) improve this task by leveraging self-attention to capture complex patterns and long range relationships between image patches. However, a key challenge for ViTs is efficiently incorporating multiscale feature representations, which is inherent in CNNs through their hierarchical structure. In this paper, we introduce the Scale-Aware Graph Attention Vision Transformer (SAG-ViT), a novel framework that addresses this challenge by integrating multi-scale features. Using EfficientNet as a backbone, the model extracts multi-scale feature maps, which are divided into patches to preserve semantic information. These patches are organized into a graph based on spatial and feature similarities, with a Graph Attention Network (GAT) refining the node embeddings. Finally, a Transformer encoder captures long-range dependencies and complex interactions. The SAG-ViT is evaluated on benchmark datasets, demonstrating its effectiveness in enhancing image classification performance. Our code and weights are publicly available at https://github.com/shravan-18/SAG-ViT
comment: 10 pages, 4 figures, 3 tables
♻ ☆ Personalized Federated Collaborative Filtering: A Variational AutoEncoder Approach
Federated Collaborative Filtering (FedCF) is an emerging field focused on developing a new recommendation framework with preserving privacy in a federated setting. Existing FedCF methods typically combine distributed Collaborative Filtering (CF) algorithms with privacy-preserving mechanisms, and then preserve personalized information into a user embedding vector. However, the user embedding is usually insufficient to preserve the rich information of the fine-grained personalization across heterogeneous clients. This paper proposes a novel personalized FedCF method by preserving users' personalized information into a latent variable and a neural model simultaneously. Specifically, we decompose the modeling of user knowledge into two encoders, each designed to capture shared knowledge and personalized knowledge separately. A personalized gating network is then applied to balance personalization and generalization between the global and local encoders. Moreover, to effectively train the proposed framework, we model the CF problem as a specialized Variational AutoEncoder (VAE) task by integrating user interaction vector reconstruction with missing value prediction. The decoder is trained to reconstruct the implicit feedback from items the user has interacted with, while also predicting items the user might be interested in but has not yet interacted with. Experimental results on benchmark datasets demonstrate that the proposed method outperforms other baseline methods, showcasing superior performance. Our code is available at https://github.com/mtics/FedDAE.
comment: 10 pages, 3 figures, 4 tables, conference
♻ ☆ A Dataset and Benchmark for Hospital Course Summarization with Adapted Large Language Models
Brief hospital course (BHC) summaries are clinical documents that summarize a patient's hospital stay. While large language models (LLMs) depict remarkable capabilities in automating real-world tasks, their capabilities for healthcare applications such as synthesizing BHCs from clinical notes have not been shown. We introduce a novel pre-processed dataset, the MIMIC-IV-BHC, encapsulating clinical note and brief hospital course (BHC) pairs to adapt LLMs for BHC synthesis. Furthermore, we introduce a benchmark of the summarization performance of two general-purpose LLMs and three healthcare-adapted LLMs. Using clinical notes as input, we apply prompting-based (using in-context learning) and fine-tuning-based adaptation strategies to three open-source LLMs (Clinical-T5-Large, Llama2-13B, FLAN-UL2) and two proprietary LLMs (GPT-3.5, GPT-4). We evaluate these LLMs across multiple context-length inputs using natural language similarity metrics. We further conduct a clinical study with five clinicians, comparing clinician-written and LLM-generated BHCs across 30 samples, focusing on their potential to enhance clinical decision-making through improved summary quality. We observe that the Llama2-13B fine-tuned LLM outperforms other domain-adapted models given quantitative evaluation metrics of BLEU and BERT-Score. GPT-4 with in-context learning shows more robustness to increasing context lengths of clinical note inputs than fine-tuned Llama2-13B. Despite comparable quantitative metrics, the reader study depicts a significant preference for summaries generated by GPT-4 with in-context learning compared to both Llama2-13B fine-tuned summaries and the original summaries, highlighting the need for qualitative clinical evaluation.
♻ ☆ DeepMDV: Learning Global Matching for Multi-depot Vehicle Routing Problems
Due to the substantial rise in online retail and e-commerce in recent years, the demand for efficient and fast solutions to Vehicle Routing Problems (VRP) has become critical. To manage the increasing demand, companies have adopted the strategy of adding more depots. However, the presence of multiple depots introduces additional complexities, making existing VRP solutions suboptimal for addressing the Multi-depot Vehicle Routing Problem (MDVRP). Traditional methods for solving the MDVRP often require significant computation time, making them unsuitable for large-scale instances. Additionally, existing learning-based solutions for the MDVRP struggle with generalizability and fail to deliver high-quality results for scenarios involving a large number of customers. In this paper, we propose a novel solution for MDVRP. Our approach employs an attention mechanism, featuring a decoder with two key layers: one layer to consider the states of all vehicles and learn to select the most suitable vehicle based on the proximity of unassigned customers, and another layer to focus on assigning a customer to the selected vehicle. This approach delivers high-quality solutions for large-scale MDVRP instances and demonstrates remarkable generalizability across varying numbers of customers and depots. Its adaptability and performance make it a practical and deployable solution for real-world logistics challenges.
♻ ☆ Extrapolated Urban View Synthesis Benchmark
Photorealistic simulators are essential for the training and evaluation of vision-centric autonomous vehicles (AVs). At their core is Novel View Synthesis (NVS), a crucial capability that generates diverse unseen viewpoints to accommodate the broad and continuous pose distribution of AVs. Recent advances in radiance fields, such as 3D Gaussian Splatting, achieve photorealistic rendering at real-time speeds and have been widely used in modeling large-scale driving scenes. However, their performance is commonly evaluated using an interpolated setup with highly correlated training and test views. In contrast, extrapolation, where test views largely deviate from training views, remains underexplored, limiting progress in generalizable simulation technology. To address this gap, we leverage publicly available AV datasets with multiple traversals, multiple vehicles, and multiple cameras to build the first Extrapolated Urban View Synthesis (EUVS) benchmark. Meanwhile, we conduct quantitative and qualitative evaluations of state-of-the-art Gaussian Splatting methods across different difficulty levels. Our results show that Gaussian Splatting is prone to overfitting to training views. Besides, incorporating diffusion priors and improving geometry cannot fundamentally improve NVS under large view changes, highlighting the need for more robust approaches and large-scale training. We have released our data to help advance self-driving and urban robotics simulation technology.
comment: Project page: https://ai4ce.github.io/EUVS-Benchmark/
♻ ☆ DCMAC: Demand-aware Customized Multi-Agent Communication via Upper Bound Training
Efficient communication can enhance the overall performance of collaborative multi-agent reinforcement learning. A common approach is to share observations through full communication, leading to significant communication overhead. Existing work attempts to perceive the global state by conducting teammate model based on local information. However, they ignore that the uncertainty generated by prediction may lead to difficult training. To address this problem, we propose a Demand-aware Customized Multi-Agent Communication (DCMAC) protocol, which use an upper bound training to obtain the ideal policy. By utilizing the demand parsing module, agent can interpret the gain of sending local message on teammate, and generate customized messages via compute the correlation between demands and local observation using cross-attention mechanism. Moreover, our method can adapt to the communication resources of agents and accelerate the training progress by appropriating the ideal policy which is trained with joint observation. Experimental results reveal that DCMAC significantly outperforms the baseline algorithms in both unconstrained and communication constrained scenarios.
comment: Paper has errors and needs to be revised and submitted
♻ ☆ M2OST: Many-to-one Regression for Predicting Spatial Transcriptomics from Digital Pathology Images AAAI 2025
The advancement of Spatial Transcriptomics (ST) has facilitated the spatially-aware profiling of gene expressions based on histopathology images. Although ST data offers valuable insights into the micro-environment of tumors, its acquisition cost remains expensive. Therefore, directly predicting the ST expressions from digital pathology images is desired. Current methods usually adopt existing regression backbones along with patch-sampling for this task, which ignores the inherent multi-scale information embedded in the pyramidal data structure of digital pathology images, and wastes the inter-spot visual information crucial for accurate gene expression prediction. To address these limitations, we propose M2OST, a many-to-one regression Transformer that can accommodate the hierarchical structure of the pathology images via a decoupled multi-scale feature extractor. Unlike traditional models that are trained with one-to-one image-label pairs, M2OST uses multiple images from different levels of the digital pathology image to jointly predict the gene expressions in their common corresponding spot. Built upon our many-to-one scheme, M2OST can be easily scaled to fit different numbers of inputs, and its network structure inherently incorporates nearby inter-spot features, enhancing regression performance. We have tested M2OST on three public ST datasets and the experimental results show that M2OST can achieve state-of-the-art performance with fewer parameters and floating-point operations (FLOPs). The code is available at: https://github.com/Dootmaan/M2OST.
comment: Accepted by AAAI 2025
♻ ☆ Doob's Lagrangian: A Sample-Efficient Variational Approach to Transition Path Sampling NeurIPS 2024
Rare event sampling in dynamical systems is a fundamental problem arising in the natural sciences, which poses significant computational challenges due to an exponentially large space of trajectories. For settings where the dynamical system of interest follows a Brownian motion with known drift, the question of conditioning the process to reach a given endpoint or desired rare event is definitively answered by Doob's h-transform. However, the naive estimation of this transform is infeasible, as it requires simulating sufficiently many forward trajectories to estimate rare event probabilities. In this work, we propose a variational formulation of Doob's h-transform as an optimization problem over trajectories between a given initial point and the desired ending point. To solve this optimization, we propose a simulation-free training objective with a model parameterization that imposes the desired boundary conditions by design. Our approach significantly reduces the search space over trajectories and avoids expensive trajectory simulation and inefficient importance sampling estimators which are required in existing methods. We demonstrate the ability of our method to find feasible transition paths on real-world molecular simulation and protein folding tasks.
comment: Accepted as Spotlight at Conference on Neural Information Processing Systems (NeurIPS 2024); Alanine dipeptide results updated after fixing unphysical parameterization and energy computation
♻ ☆ Bi-level Contrastive Learning for Knowledge-Enhanced Molecule Representations AAAI 2025
Molecular representation learning is vital for various downstream applications, including the analysis and prediction of molecular properties and side effects. While Graph Neural Networks (GNNs) have been a popular framework for modeling molecular data, they often struggle to capture the full complexity of molecular representations. In this paper, we introduce a novel method called GODE, which accounts for the dual-level structure inherent in molecules. Molecules possess an intrinsic graph structure and simultaneously function as nodes within a broader molecular knowledge graph. GODE integrates individual molecular graph representations with multi-domain biochemical data from knowledge graphs. By pre-training two GNNs on different graph structures and employing contrastive learning, GODE effectively fuses molecular structures with their corresponding knowledge graph substructures. This fusion yields a more robust and informative representation, enhancing molecular property predictions by leveraging both chemical and biological information. When fine-tuned across 11 chemical property tasks, our model significantly outperforms existing benchmarks, achieving an average ROC-AUC improvement of 12.7% for classification tasks and an average RMSE/MAE improvement of 34.4% for regression tasks. Notably, GODE surpasses the current leading model in property prediction, with advancements of 2.2% in classification and 7.2% in regression tasks.
comment: AAAI 2025
♻ ☆ Gradient Diffusion: A Perturbation-Resilient Gradient Leakage Attack AAAI 2025
Recent years have witnessed the vulnerability of Federated Learning (FL) against gradient leakage attacks, where the private training data can be recovered from the exchanged gradients, making gradient protection a critical issue for the FL training process. Existing solutions often resort to perturbation-based mechanisms, such as differential privacy, where each participating client injects a specific amount of noise into local gradients before aggregating to the server, and the global distribution variation finally conceals the gradient privacy. However, perturbation is not always the panacea for gradient protection since the robustness heavily relies on the injected noise. This intuition raises an interesting question: \textit{is it possible to deactivate existing protection mechanisms by removing the perturbation inside the gradients?} In this paper, we present the answer: \textit{yes} and propose the Perturbation-resilient Gradient Leakage Attack (PGLA), the first attempt to recover the perturbed gradients, without additional access to the original model structure or third-party data. Specifically, we leverage the inherent diffusion property of gradient perturbation protection and construct a novel diffusion-based denoising model to implement PGLA. Our insight is that capturing the disturbance level of perturbation during the diffusion reverse process can release the gradient denoising capability, which promotes the diffusion model to generate approximate gradients as the original clean version through adaptive sampling steps. Extensive experiments demonstrate that PGLA effectively recovers the protected gradients and exposes the FL training process to the threat of gradient leakage, achieving the best quality in gradient denoising and data recovery compared to existing models. We hope to arouse public attention on PGLA and its defense.
comment: Accepted by AAAI 2025
Computation and Language 109
☆ Delve into Visual Contrastive Decoding for Hallucination Mitigation of Large Vision-Language Models
While large vision-language models (LVLMs) have shown impressive capabilities in generating plausible responses correlated with input visual contents, they still suffer from hallucinations, where the generated text inaccurately reflects visual contents. To address this, recent approaches apply contrastive decoding to calibrate the model's response via contrasting output distributions with original and visually distorted samples, demonstrating promising hallucination mitigation in a training-free manner. However, the potential of changing information in visual inputs is not well-explored, so a deeper investigation into the behaviors of visual contrastive decoding is of great interest. In this paper, we first explore various methods for contrastive decoding to change visual contents, including image downsampling and editing. Downsampling images reduces the detailed textual information while editing yields new contents in images, providing new aspects as visual contrastive samples. To further study benefits by using different contrastive samples, we analyze probability-level metrics, including entropy and distribution distance. Interestingly, the effect of these samples in mitigating hallucinations varies a lot across LVLMs and benchmarks. Based on our analysis, we propose a simple yet effective method to combine contrastive samples, offering a practical solution for applying contrastive decoding across various scenarios. Extensive experiments are conducted to validate the proposed fusion method among different benchmarks.
comment: Under review. Project pages: https://github.com/YiLunLee/VCD_Analysis
☆ Training Large Language Models to Reason in a Continuous Latent Space
Large language models (LLMs) are restricted to reason in the "language space", where they typically express the reasoning process with a chain-of-thought (CoT) to solve a complex reasoning problem. However, we argue that language space may not always be optimal for reasoning. For example, most word tokens are primarily for textual coherence and not essential for reasoning, while some critical tokens require complex planning and pose huge challenges to LLMs. To explore the potential of LLM reasoning in an unrestricted latent space instead of using natural language, we introduce a new paradigm Coconut (Chain of Continuous Thought). We utilize the last hidden state of the LLM as a representation of the reasoning state (termed "continuous thought"). Rather than decoding this into a word token, we feed it back to the LLM as the subsequent input embedding directly in the continuous space. Experiments show that Coconut can effectively augment the LLM on several reasoning tasks. This novel latent reasoning paradigm leads to emergent advanced reasoning patterns: the continuous thought can encode multiple alternative next reasoning steps, allowing the model to perform a breadth-first search (BFS) to solve the problem, rather than prematurely committing to a single deterministic path like CoT. Coconut outperforms CoT in certain logical reasoning tasks that require substantial backtracking during planning, with fewer thinking tokens during inference. These findings demonstrate the promise of latent reasoning and offer valuable insights for future research.
☆ Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
A key component of building safe and reliable language models is enabling the models to appropriately refuse to follow certain instructions or answer certain questions. We may want models to output refusal messages for various categories of user queries, for example, ill-posed questions, instructions for committing illegal acts, or queries which require information past the model's knowledge horizon. Engineering models that refuse to answer such questions is complicated by the fact that an individual may want their model to exhibit varying levels of sensitivity for refusing queries of various categories, and different users may want different refusal rates. The current default approach involves training multiple models with varying proportions of refusal messages from each category to achieve the desired refusal rates, which is computationally expensive and may require training a new model to accommodate each user's desired preference over refusal rates. To address these challenges, we propose refusal tokens, one such token for each refusal category or a single refusal token, which are prepended to the model's responses during training. We then show how to increase or decrease the probability of generating the refusal token for each category during inference to steer the model's refusal behavior. Refusal tokens enable controlling a single model's refusal rates without the need of any further fine-tuning, but only by selectively intervening during generation.
comment: 19 pages
☆ ONEBench to Test Them All: Sample-Level Benchmarking Over Open-Ended Capabilities
Traditional fixed test sets fall short in evaluating open-ended capabilities of foundation models. To address this, we propose ONEBench(OpeN-Ended Benchmarking), a new testing paradigm that consolidates individual evaluation datasets into a unified, ever-expanding sample pool. ONEBench allows users to generate custom, open-ended evaluation benchmarks from this pool, corresponding to specific capabilities of interest. By aggregating samples across test sets, ONEBench enables the assessment of diverse capabilities beyond those covered by the original test sets, while mitigating overfitting and dataset bias. Most importantly, it frames model evaluation as a collective process of selecting and aggregating sample-level tests. The shift from task-specific benchmarks to ONEBench introduces two challenges: (1)heterogeneity and (2)incompleteness. Heterogeneity refers to the aggregation over diverse metrics, while incompleteness describes comparing models evaluated on different data subsets. To address these challenges, we explore algorithms to aggregate sparse measurements into reliable model scores. Our aggregation algorithm ensures identifiability(asymptotically recovering ground-truth scores) and rapid convergence, enabling accurate model ranking with less data. On homogenous datasets, we show our aggregation algorithm provides rankings that highly correlate with those produced by average scores. We also demonstrate robustness to ~95% of measurements missing, reducing evaluation cost by up to 20x with little-to-no change in model rankings. We introduce ONEBench-LLM for language models and ONEBench-LMM for vision-language models, unifying evaluations across these domains. Overall, we present a technique for open-ended evaluation, which can aggregate over incomplete, heterogeneous sample-level measurements to continually grow a benchmark alongside the rapidly developing foundation models.
☆ JAPAGEN: Efficient Few/Zero-shot Learning via Japanese Training Dataset Generation with LLM ACL
Recently some studies have highlighted the potential of Large Language Models (LLMs) as effective generators of supervised training data, offering advantages such as enhanced inference efficiency and reduced costs associated with data collection. However, these studies have predominantly focused on English language tasks. In this paper, we address the fundamental research question: Can LLMs serve as proficient training data generators for other language tasks? Specifically, we leverage LLMs to synthesize supervised training data under few-shot and zero-shot learning scenarios across six diverse Japanese downstream tasks. Subsequently, we utilize this synthesized data to train compact models (e.g., BERT). This novel methodology is termed JAPAGEN. Our experimental findings underscore that JAPAGEN achieves robust performance in classification tasks that necessitate formal text inputs, demonstrating competitive results compared to conventional LLM prompting strategies.
comment: Accepted by PACLIC38 (2024)
☆ AutoDCWorkflow: LLM-based Data Cleaning Workflow Auto-Generation and Benchmark
We investigate the reasoning capabilities of large language models (LLMs) for automatically generating data-cleaning workflows. To evaluate LLMs' ability to complete data-cleaning tasks, we implemented a pipeline for LLM-based Auto Data Cleaning Workflow (AutoDCWorkflow), prompting LLMs on data cleaning operations to repair three types of data quality issues: duplicates, missing values, and inconsistent data formats. Given a dirty table and a purpose (expressed as a query), this pipeline generates a minimal, clean table sufficient to address the purpose and the data cleaning workflow used to produce the table. The planning process involves three main LLM-driven components: (1) Select Target Columns: Identifies a set of target columns related to the purpose. (2) Inspect Column Quality: Assesses the data quality for each target column and generates a Data Quality Report as operation objectives. (3) Generate Operation & Arguments: Predicts the next operation and arguments based on the data quality report results. Additionally, we propose a data cleaning benchmark to evaluate the capability of LLM agents to automatically generate workflows that address data cleaning purposes of varying difficulty levels. The benchmark comprises the annotated datasets as a collection of purpose, raw table, clean table, data cleaning workflow, and answer set. In our experiments, we evaluated three LLMs that auto-generate purpose-driven data cleaning workflows. The results indicate that LLMs perform well in planning and generating data-cleaning workflows without the need for fine-tuning.
☆ VP-MEL: Visual Prompts Guided Multimodal Entity Linking
Multimodal Entity Linking (MEL) is extensively utilized in the domains of information retrieval. However, existing MEL methods typically utilize mention words as mentions for retrieval. This results in a significant dependence of MEL on mention words, thereby constraining its capacity to effectively leverage information from both images and text. In situations where mention words are absent, MEL methods struggle to leverage image-text pairs for entity linking. To solve these issues, we introduce a Visual Prompts guided Multimodal Entity Linking (VP-MEL) task. VP-MEL directly marks specific regions within the image. These markers are referred to as visual prompts in VP-MEL. Without mention words, VP-MEL aims to utilize marked image-text pairs to align visual prompts with specific entities in the knowledge bases. A new dataset for the VP-MEL task, VPWiki, is proposed in this paper. Moreover, we propose a framework named FBMEL, which enhances the significance of visual prompts and fully leverages the information in image-text pairs. Experimental results on the VPWiki dataset demonstrate that FBMEL outperforms baseline methods across multiple benchmarks for the VP-MEL task.
☆ How to Merge Your Multimodal Models Over Time?
Model merging combines multiple expert models - finetuned from a base foundation model on diverse tasks and domains - into a single, more capable model. However, most existing model merging approaches assume that all experts are available simultaneously. In reality, new tasks and domains emerge progressively over time, requiring strategies to integrate the knowledge of expert models as they become available: a process we call temporal model merging. The temporal dimension introduces unique challenges not addressed in prior work, raising new questions such as: when training for a new task, should the expert model start from the merged past experts or from the original base model? Should we merge all models at each time step? Which merging techniques are best suited for temporal merging? Should different strategies be used to initialize the training and deploy the model? To answer these questions, we propose a unified framework called TIME - Temporal Integration of Model Expertise - which defines temporal model merging across three axes: (1) Initialization Phase, (2) Deployment Phase, and (3) Merging Technique. Using TIME, we study temporal model merging across model sizes, compute budgets, and learning horizons on the FoMo-in-Flux benchmark. Our comprehensive suite of experiments across TIME allows us to uncover key insights for temporal model merging, offering a better understanding of current challenges and best practices for effective temporal model merging.
comment: Technical Report. Code at https://github.com/ExplainableML/fomo_in_flux
☆ OmniEvalKit: A Modular, Lightweight Toolbox for Evaluating Large Language Model and its Omni-Extensions
The rapid advancements in Large Language Models (LLMs) have significantly expanded their applications, ranging from multilingual support to domain-specific tasks and multimodal integration. In this paper, we present OmniEvalKit, a novel benchmarking toolbox designed to evaluate LLMs and their omni-extensions across multilingual, multidomain, and multimodal capabilities. Unlike existing benchmarks that often focus on a single aspect, OmniEvalKit provides a modular, lightweight, and automated evaluation system. It is structured with a modular architecture comprising a Static Builder and Dynamic Data Flow, promoting the seamless integration of new models and datasets. OmniEvalKit supports over 100 LLMs and 50 evaluation datasets, covering comprehensive evaluations across thousands of model-dataset combinations. OmniEvalKit is dedicated to creating an ultra-lightweight and fast-deployable evaluation framework, making downstream applications more convenient and versatile for the AI community.
☆ I Don't Know: Explicit Modeling of Uncertainty with an [IDK] Token NeurIPS 2024
Large Language Models are known to capture real-world knowledge, allowing them to excel in many downstream tasks. Despite recent advances, these models are still prone to what are commonly known as hallucinations, causing them to emit unwanted and factually incorrect text. In this work, we propose a novel calibration method that can be used to combat hallucinations. We add a special [IDK] ("I don't know") token to the model's vocabulary and introduce an objective function that shifts probability mass to the [IDK] token for incorrect predictions. This approach allows the model to express uncertainty in its output explicitly. We evaluate our proposed method across multiple model architectures and factual downstream tasks. We find that models trained with our method are able to express uncertainty in places where they would previously make mistakes while suffering only a small loss of encoded knowledge. We further perform extensive ablation studies of multiple variations of our approach and provide a detailed analysis of the precision-recall tradeoff of our method.
comment: Published at NeurIPS 2024
☆ GEAR: A Simple GENERATE, EMBED, AVERAGE AND RANK Approach for Unsupervised Reverse Dictionary COLING 2025
Reverse Dictionary (RD) is the task of obtaining the most relevant word or set of words given a textual description or dictionary definition. Effective RD methods have applications in accessibility, translation or writing support systems. Moreover, in NLP research we find RD to be used to benchmark text encoders at various granularities, as it often requires word, definition and sentence embeddings. In this paper, we propose a simple approach to RD that leverages LLMs in combination with embedding models. Despite its simplicity, this approach outperforms supervised baselines in well studied RD datasets, while also showing less over-fitting. We also conduct a number of experiments on different dictionaries and analyze how different styles, registers and target audiences impact the quality of RD systems. We conclude that, on average, untuned embeddings alone fare way below an LLM-only baseline (although they are competitive in highly technical dictionaries), but are crucial for boosting performance in combined methods.
comment: 9 pages, Accepted at COLING 2025
☆ Copyright-Protected Language Generation via Adaptive Model Fusion
The risk of language models reproducing copyrighted material from their training data has led to the development of various protective measures. Among these, inference-time strategies that impose constraints via post-processing have shown promise in addressing the complexities of copyright regulation. However, they often incur prohibitive computational costs or suffer from performance trade-offs. To overcome these limitations, we introduce Copyright-Protecting Model Fusion (CP-Fuse), a novel approach that combines models trained on disjoint sets of copyrighted material during inference. In particular, CP-Fuse adaptively aggregates the model outputs to minimize the reproduction of copyrighted content, adhering to a crucial balancing property that prevents the regurgitation of memorized data. Through extensive experiments, we show that CP-Fuse significantly reduces the reproduction of protected material without compromising the quality of text and code generation. Moreover, its post-hoc nature allows seamless integration with other protective measures, further enhancing copyright safeguards. Lastly, we show that CP-Fuse is robust against common techniques for extracting training data.
comment: 47 pages, 21 Figures. arXiv admin note: substantial text overlap with arXiv:2407.20105
☆ Towards Controllable Speech Synthesis in the Era of Large Language Models: A Survey
Text-to-speech (TTS), also known as speech synthesis, is a prominent research area that aims to generate natural-sounding human speech from text. Recently, with the increasing industrial demand, TTS technologies have evolved beyond synthesizing human-like speech to enabling controllable speech generation. This includes fine-grained control over various attributes of synthesized speech such as emotion, prosody, timbre, and duration. Besides, advancements in deep learning, such as diffusion and large language models, have significantly enhanced controllable TTS over the past several years. In this paper, we conduct a comprehensive survey of controllable TTS, covering approaches ranging from basic control techniques to methods utilizing natural language prompts, aiming to provide a clear understanding of the current state of research. We examine the general controllable TTS pipeline, challenges, model architectures, and control strategies, offering a comprehensive and clear taxonomy of existing methods. Additionally, we provide a detailed summary of datasets and evaluation metrics and shed some light on the applications and future directions of controllable TTS. To the best of our knowledge, this survey paper provides the first comprehensive review of emerging controllable TTS methods, which can serve as a beneficial resource for both academic researchers and industry practitioners.
comment: A comprehensive survey on controllable TTS, 23 pages, 6 tables, 4 figures, 280 references
☆ Anchoring Bias in Large Language Models: An Experimental Study
Large Language Models (LLMs) like GPT-4 and Gemini have significantly advanced artificial intelligence by enabling machines to generate and comprehend human-like text. Despite their impressive capabilities, LLMs are not immune to limitations, including various biases. While much research has explored demographic biases, the cognitive biases in LLMs have not been equally scrutinized. This study delves into anchoring bias, a cognitive bias where initial information disproportionately influences judgment. Utilizing an experimental dataset, we examine how anchoring bias manifests in LLMs and verify the effectiveness of various mitigation strategies. Our findings highlight the sensitivity of LLM responses to biased hints. At the same time, our experiments show that, to mitigate anchoring bias, one needs to collect hints from comprehensive angles to prevent the LLMs from being anchored to individual pieces of information, while simple algorithms such as Chain-of-Thought, Thoughts of Principles, Ignoring Anchor Hints, and Reflection are not sufficient.
☆ Data Quality Enhancement on the Basis of Diversity with Large Language Models for Text Classification: Uncovered, Difficult, and Noisy COLING 2025
In recent years, the use of large language models (LLMs) for text classification has attracted widespread attention. Despite this, the classification accuracy of LLMs has not yet universally surpassed that of smaller models. LLMs can enhance their performance in text classification through fine-tuning. However, existing data quality research based on LLMs is challenging to apply directly to solve text classification problems. To further improve the performance of LLMs in classification tasks, this paper proposes a data quality enhancement (DQE) method for text classification based on LLMs. This method starts by using a greedy algorithm to select data, dividing the dataset into sampled and unsampled subsets, and then performing fine-tuning of the LLMs using the sampled data. Subsequently, this model is used to predict the outcomes for the unsampled data, categorizing incorrectly predicted data into uncovered, difficult, and noisy data. Experimental results demonstrate that our method effectively enhances the performance of LLMs in text classification tasks and significantly improves training efficiency, saving nearly half of the training time. Our method has achieved state-of-the-art performance in several open-source classification tasks.
comment: Accepted by COLING 2025(main, long paper)
☆ ProcessBench: Identifying Process Errors in Mathematical Reasoning
As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.
☆ Understanding Factual Recall in Transformers via Associative Memories
Large language models have demonstrated an impressive ability to perform factual recall. Prior work has found that transformers trained on factual recall tasks can store information at a rate proportional to their parameter count. In our work, we show that shallow transformers can use a combination of associative memories to obtain such near optimal storage capacity. We begin by proving that the storage capacities of both linear and MLP associative memories scale linearly with parameter count. We next introduce a synthetic factual recall task, and prove that a transformer with a single layer of self-attention followed by an MLP can obtain 100% accuracy on the task whenever either the total number of self-attention parameters or MLP parameters scales (up to log factors) linearly with the number of facts. In particular, the transformer can trade off between using the value matrices or the MLP as an associative memory to store the dataset of facts. We complement these expressivity results with an analysis of the gradient flow trajectory of a simplified linear attention model trained on our factual recall task, where we show that the model exhibits sequential learning behavior.
☆ The Fusion of Large Language Models and Formal Methods for Trustworthy AI Agents: A Roadmap
Large Language Models (LLMs) have emerged as a transformative AI paradigm, profoundly influencing daily life through their exceptional language understanding and contextual generation capabilities. Despite their remarkable performance, LLMs face a critical challenge: the propensity to produce unreliable outputs due to the inherent limitations of their learning-based nature. Formal methods (FMs), on the other hand, are a well-established computation paradigm that provides mathematically rigorous techniques for modeling, specifying, and verifying the correctness of systems. FMs have been extensively applied in mission-critical software engineering, embedded systems, and cybersecurity. However, the primary challenge impeding the deployment of FMs in real-world settings lies in their steep learning curves, the absence of user-friendly interfaces, and issues with efficiency and adaptability. This position paper outlines a roadmap for advancing the next generation of trustworthy AI systems by leveraging the mutual enhancement of LLMs and FMs. First, we illustrate how FMs, including reasoning and certification techniques, can help LLMs generate more reliable and formally certified outputs. Subsequently, we highlight how the advanced learning capabilities and adaptability of LLMs can significantly enhance the usability, efficiency, and scalability of existing FM tools. Finally, we show that unifying these two computation paradigms -- integrating the flexibility and intelligence of LLMs with the rigorous reasoning abilities of FMs -- has transformative potential for the development of trustworthy AI software systems. We acknowledge that this integration has the potential to enhance both the trustworthiness and efficiency of software engineering practices while fostering the development of intelligent FM tools capable of addressing complex yet real-world challenges.
comment: 24 pages, 4 figures
☆ Small Languages, Big Models: A Study of Continual Training on Languages of Norway
Training large language models requires vast amounts of data, posing a challenge for less widely spoken languages like Norwegian and even more so for truly low-resource languages like S\'ami. To address this issue, we present a novel three-stage continual training approach. We also experiment with combining causal and masked language modeling to get more flexible models. Based on our findings, we train, evaluate, and openly release a new large generative language model for Norwegian Bokm\r{a}l, Nynorsk, and Northern S\'ami with 11.4 billion parameters: NorMistral-11B.
comment: pre-print, under review
☆ SafeWorld: Geo-Diverse Safety Alignment NeurIPS 2024
In the rapidly evolving field of Large Language Models (LLMs), ensuring safety is a crucial and widely discussed topic. However, existing works often overlook the geo-diversity of cultural and legal standards across the world. To demonstrate the challenges posed by geo-diverse safety standards, we introduce SafeWorld, a novel benchmark specifically designed to evaluate LLMs' ability to generate responses that are not only helpful but also culturally sensitive and legally compliant across diverse global contexts. SafeWorld encompasses 2,342 test user queries, each grounded in high-quality, human-verified cultural norms and legal policies from 50 countries and 493 regions/races. On top of it, we propose a multi-dimensional automatic safety evaluation framework that assesses the contextual appropriateness, accuracy, and comprehensiveness of responses. Our evaluations reveal that current LLMs struggle to meet these criteria. To enhance LLMs' alignment with geo-diverse safety standards, we synthesize helpful preference pairs for Direct Preference Optimization (DPO) alignment training. The preference pair construction aims to encourage LLMs to behave appropriately and provide precise references to relevant cultural norms and policies when necessary. Our trained SafeWorldLM outperforms all competing models, including GPT-4o on all three evaluation dimensions by a large margin. Global human evaluators also note a nearly 20% higher winning rate in helpfulness and harmfulness evaluation. Our code and data can be found here: https://github.com/PlusLabNLP/SafeWorld.
comment: Accepted by NeurIPS 2024
☆ Gated Delta Networks: Improving Mamba2 with Delta Rule
Linear Transformers have gained attention as efficient alternatives to standard Transformers, but their performance in retrieval and long-context tasks has been limited. To address these limitations, recent work has explored two distinct mechanisms: gating for adaptive memory control and the delta update rule for precise memory modifications. We observe that these mechanisms are complementary: gating enables rapid memory erasure while the delta rule facilitates targeted updates. Building on this insight, we introduce the gated delta rule and develop a parallel training algorithm optimized for modern hardware. Our proposed architecture, Gated DeltaNet, consistently surpasses existing models like Mamba2 and DeltaNet across multiple benchmarks, including language modeling, common-sense reasoning, in-context retrieval, length extrapolation, and long-context understanding. We further enhance performance by developing hybrid architectures that combine Gated DeltaNet layers with sliding window attention or Mamba2 layers, achieving both improved training efficiency and superior task performance.
comment: Preprint
☆ BoRA: Bi-dimensional Weight-Decomposed Low-Rank Adaptation
In recent years, Parameter-Efficient Fine-Tuning (PEFT) methods like Low-Rank Adaptation (LoRA) have significantly enhanced the adaptability of large-scale pre-trained models. Weight-Decomposed Low-Rank Adaptation (DoRA) improves upon LoRA by separating the magnitude and direction components of the weight matrix, leading to superior performance. However, DoRA's improvements are limited to the vertical dimension, resulting in an asymmetrical pattern between horizontal and vertical dimensions. This paper introduces BoRA, an innovative extension of LoRA and DoRA, characterized by symmetrical properties across horizontal and vertical dimensions. Our approach optimizes the weight matrix symmetrically by adjusting both column-wise and row-wise magnitudes. Extensive experiments demonstrate that BoRA surpasses state-of-the-art PEFT methods, including LoRA and DoRA, achieving superior results across various benchmarks.
☆ Integrating Expert Labels into LLM-based Emission Goal Detection: Example Selection vs Automatic Prompt Design
We address the detection of emission reduction goals in corporate reports, an important task for monitoring companies' progress in addressing climate change. Specifically, we focus on the issue of integrating expert feedback in the form of labeled example passages into LLM-based pipelines, and compare the two strategies of (1) a dynamic selection of few-shot examples and (2) the automatic optimization of the prompt by the LLM itself. Our findings on a public dataset of 769 climate-related passages from real-world business reports indicate that automatic prompt optimization is the superior approach, while combining both methods provides only limited benefit. Qualitative results indicate that optimized prompts do indeed capture many intricacies of the targeted emission goal extraction task.
LLM-BIP: Structured Pruning for Large Language Models with Block-Wise Forward Importance Propagation
Large language models (LLMs) have demonstrated remarkable performance across various language tasks, but their widespread deployment is impeded by their large size and high computational costs. Structural pruning is a prevailing technique used to introduce sparsity into pre-trained models and facilitate direct hardware acceleration during inference by removing redundant connections (structurally-grouped parameters), such as channels and attention heads. Existing structural pruning approaches often employ either global or layer-wise pruning criteria; however, they are hindered by ineffectiveness stemming from inaccurate evaluation of connection importance. Global pruning methods typically assess component importance using near-zero and unreliable gradients, while layer-wise pruning approaches encounter significant pruning error accumulation issues. To this end, we propose a more accurate pruning metric based on the block-wise importance score propagation, termed LLM-BIP. Specifically, LLM-BIP precisely evaluates connection importance by gauging its influence on the respective transformer block output, which can be efficiently approximated in a single forward pass through an upper bound derived from the assumption of Lipschitz continuity. We evaluate the proposed method using LLaMA-7B, Vicuna-7B, and LLaMA-13B across common zero-shot tasks. The results demonstrate that our approach achieves an average of 3.26% increase in accuracy for common reasoning tasks compared to previous best baselines. It also reduces perplexity by 14.09 and 68.76 on average for the WikiText2 dataset and PTB dataset, respectively.
☆ StarWhisper Telescope: Agent-Based Observation Assistant System to Approach AI Astrophysicist
With the rapid advancements in Large Language Models (LLMs), LLM-based agents have introduced convenient and user-friendly methods for leveraging tools across various domains. In the field of astronomical observation, the construction of new telescopes has significantly increased astronomers' workload. Deploying LLM-powered agents can effectively alleviate this burden and reduce the costs associated with training personnel. Within the Nearby Galaxy Supernovae Survey (NGSS) project, which encompasses eight telescopes across three observation sites, aiming to find the transients from the galaxies in 50 mpc, we have developed the \textbf{StarWhisper Telescope System} to manage the entire observation process. This system automates tasks such as generating observation lists, conducting observations, analyzing data, and providing feedback to the observer. Observation lists are customized for different sites and strategies to ensure comprehensive coverage of celestial objects. After manual verification, these lists are uploaded to the telescopes via the agents in the system, which initiates observations upon neutral language. The observed images are analyzed in real-time, and the transients are promptly communicated to the observer. The agent modifies them into a real-time follow-up observation proposal and send to the Xinglong observatory group chat, then add them to the next-day observation lists. Additionally, the integration of AI agents within the system provides online accessibility, saving astronomers' time and encouraging greater participation from amateur astronomers in the NGSS project.
comment: 21 pages, 18 figures
☆ GameArena: Evaluating LLM Reasoning through Live Computer Games
Evaluating the reasoning abilities of large language models (LLMs) is challenging. Existing benchmarks often depend on static datasets, which are vulnerable to data contamination and may get saturated over time, or on binary live human feedback that conflates reasoning with other abilities. As the most prominent dynamic benchmark, Chatbot Arena evaluates open-ended questions in real-world settings, but lacks the granularity in assessing specific reasoning capabilities. We introduce GameArena, a dynamic benchmark designed to evaluate LLM reasoning capabilities through interactive gameplay with humans. GameArena consists of three games designed to test specific reasoning capabilities (e.g., deductive and inductive reasoning), while keeping participants entertained and engaged. We analyze the gaming data retrospectively to uncover the underlying reasoning processes of LLMs and measure their fine-grained reasoning capabilities. We collect over 2000 game sessions and provide detailed assessments of various reasoning capabilities for five state-of-the-art LLMs. Our user study with 100 participants suggests that GameArena improves user engagement compared to Chatbot Arena. For the first time, GameArena enables the collection of step-by-step LLM reasoning data in the wild.
☆ Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection SC
Automatic Speech Recognition (ASR) plays an important role in speech-based automatic detection of Alzheimer's disease (AD). However, recognition errors could propagate downstream, potentially impacting the detection decisions. Recent studies have revealed a non-linear relationship between word error rates (WER) and AD detection performance, where ASR transcriptions with notable errors could still yield AD detection accuracy equivalent to that based on manual transcriptions. This work presents a series of analyses to explore the effect of ASR transcription errors in BERT-based AD detection systems. Our investigation reveals that not all ASR errors contribute equally to detection performance. Certain words, such as stopwords, despite constituting a large proportion of errors, are shown to play a limited role in distinguishing AD. In contrast, the keywords related to diagnosis tasks exhibit significantly greater importance relative to other words. These findings provide insights into the interplay between ASR errors and the downstream detection model.
comment: Accepted by IEEE ISCSLP 2024
☆ PediaBench: A Comprehensive Chinese Pediatric Dataset for Benchmarking Large Language Models
The emergence of Large Language Models (LLMs) in the medical domain has stressed a compelling need for standard datasets to evaluate their question-answering (QA) performance. Although there have been several benchmark datasets for medical QA, they either cover common knowledge across different departments or are specific to another department rather than pediatrics. Moreover, some of them are limited to objective questions and do not measure the generation capacity of LLMs. Therefore, they cannot comprehensively assess the QA ability of LLMs in pediatrics. To fill this gap, we construct PediaBench, the first Chinese pediatric dataset for LLM evaluation. Specifically, it contains 4,565 objective questions and 1,632 subjective questions spanning 12 pediatric disease groups. It adopts an integrated scoring criterion based on different difficulty levels to thoroughly assess the proficiency of an LLM in instruction following, knowledge understanding, clinical case analysis, etc. Finally, we validate the effectiveness of PediaBench with extensive experiments on 20 open-source and commercial LLMs. Through an in-depth analysis of experimental results, we offer insights into the ability of LLMs to answer pediatric questions in the Chinese context, highlighting their limitations for further improvements. Our code and data are published at https://github.com/ACMISLab/PediaBench.
comment: 21 pages, 12 figures
☆ Methods for Legal Citation Prediction in the Age of LLMs: An Australian Law Case Study
In recent years, Large Language Models (LLMs) have shown great potential across a wide range of legal tasks. Despite these advances, mitigating hallucination remains a significant challenge, with state-of-the-art LLMs still frequently generating incorrect legal references. In this paper, we focus on the problem of legal citation prediction within the Australian law context, where correctly identifying and citing relevant legislations or precedents is critical. We compare several approaches: prompting general purpose and law-specialised LLMs, retrieval-only pipelines with both generic and domain-specific embeddings, task-specific instruction-tuning of LLMs, and hybrid strategies that combine LLMs with retrieval augmentation, query expansion, or voting ensembles. Our findings indicate that domain-specific pre-training alone is insufficient for achieving satisfactory citation accuracy even after law-specialised pre-training. In contrast, instruction tuning on our task-specific dataset dramatically boosts performance reaching the best results across all settings. We also highlight that database granularity along with the type of embeddings play a critical role in the performance of retrieval systems. Among retrieval-based approaches, hybrid methods consistently outperform retrieval-only setups, and among these, ensemble voting delivers the best result by combining the predictive quality of instruction-tuned LLMs with the retrieval system.
comment: For code, data, and models see https://auslawbench.github.io
☆ Optimizing Multi-Task Learning for Enhanced Performance in Large Language Models
This study aims to explore the performance improvement method of large language models based on GPT-4 under the multi-task learning framework and conducts experiments on two tasks: text classification and automatic summary generation. Through the combined design of shared feature extractors and task-specific modules, we achieve knowledge-sharing and optimization of multiple tasks in the same model. The experiment uses multiple subtasks of the GLUE dataset to compare the performance of the multi-task model with the single-task GPT-4, the multi-task version of GPT-3, the BERT basic model, and the classic Bi-LSTM with Attention model. The results show that the proposed multi-task learning model outperforms other comparison models in terms of text classification accuracy and ROUGE value of summary generation, demonstrating the advantages of multi-task learning in improving model generalization ability and collaborative learning between tasks. The model maintains a stable loss convergence rate during training, showing good learning efficiency and adaptability to the test set. This study verifies the applicability of the multi-task learning framework in large language models, especially in improving the model's ability to balance different tasks. In the future, with the combination of large language models and multimodal data and the application of dynamic task adjustment technology, the framework based on multi-task learning is expected to play a greater role in practical applications across fields and provide new ideas for the development of general artificial intelligence.
☆ A Comparative Study of Learning Paradigms in Large Language Models via Intrinsic Dimension
The performance of Large Language Models (LLMs) on natural language tasks can be improved through both supervised fine-tuning (SFT) and in-context learning (ICL), which operate via distinct mechanisms. Supervised fine-tuning updates the model's weights by minimizing loss on training data, whereas in-context learning leverages task demonstrations embedded in the prompt, without changing the model's parameters. This study investigates the effects of these learning paradigms on the hidden representations of LLMs using Intrinsic Dimension (ID). We use ID to estimate the number of degrees of freedom between representations extracted from LLMs as they perform specific natural language tasks. We first explore how the ID of LLM representations evolves during SFT and how it varies due to the number of demonstrations in ICL. We then compare the IDs induced by SFT and ICL and find that ICL consistently induces a higher ID compared to SFT, suggesting that representations generated during ICL reside in higher dimensional manifolds in the embedding space.
LLMs as Debate Partners: Utilizing Genetic Algorithms and Adversarial Search for Adaptive Arguments
This paper introduces DebateBrawl, an innovative AI-powered debate platform that integrates Large Language Models (LLMs), Genetic Algorithms (GA), and Adversarial Search (AS) to create an adaptive and engaging debating experience. DebateBrawl addresses the limitations of traditional LLMs in strategic planning by incorporating evolutionary optimization and game-theoretic techniques. The system demonstrates remarkable performance in generating coherent, contextually relevant arguments while adapting its strategy in real-time. Experimental results involving 23 debates show balanced outcomes between AI and human participants, with the AI system achieving an average score of 2.72 compared to the human average of 2.67 out of 10. User feedback indicates significant improvements in debating skills and a highly satisfactory learning experience, with 85% of users reporting improved debating abilities and 78% finding the AI opponent appropriately challenging. The system's ability to maintain high factual accuracy (92% compared to 78% in human-only debates) while generating diverse arguments addresses critical concerns in AI-assisted discourse. DebateBrawl not only serves as an effective educational tool but also contributes to the broader goal of improving public discourse through AI-assisted argumentation. The paper discusses the ethical implications of AI in persuasive contexts and outlines the measures implemented to ensure responsible development and deployment of the system, including robust fact-checking mechanisms and transparency in decision-making processes.
☆ SiReRAG: Indexing Similar and Related Information for Multihop Reasoning
Indexing is an important step towards strong performance in retrieval-augmented generation (RAG) systems. However, existing methods organize data based on either semantic similarity (similarity) or related information (relatedness), but do not cover both perspectives comprehensively. Our analysis reveals that modeling only one perspective results in insufficient knowledge synthesis, leading to suboptimal performance on complex tasks requiring multihop reasoning. In this paper, we propose SiReRAG, a novel RAG indexing approach that explicitly considers both similar and related information. On the similarity side, we follow existing work and explore some variances to construct a similarity tree based on recursive summarization. On the relatedness side, SiReRAG extracts propositions and entities from texts, groups propositions via shared entities, and generates recursive summaries to construct a relatedness tree. We index and flatten both similarity and relatedness trees into a unified retrieval pool. Our experiments demonstrate that SiReRAG consistently outperforms state-of-the-art indexing methods on three multihop datasets (MuSiQue, 2WikiMultiHopQA, and HotpotQA), with an average 1.9% improvement in F1 scores. As a reasonably efficient solution, SiReRAG enhances existing reranking methods significantly, with up to 7.8% improvement in average F1 scores.
☆ SparseAccelerate: Efficient Long-Context Inference for Mid-Range GPUs
As Large Language Models (LLMs) scale to longer context windows, the computational cost of attention mechanisms, which traditionally grows quadratically with input length, presents a critical challenge for real-time and memory-constrained deployments. Existing sparse attention techniques have sought to reduce this complexity, but they often incur significant overhead or compromise accuracy, making them less practical for large contexts on mid-range hardware. In this paper, we introduce SparseAccelerate, a dynamic sparse attention method that adapts its sparsity patterns based on input characteristics, effectively flattening the attention complexity curve. Our approach is effective for input lengths starting at 16K tokens and scales efficiently up to 128K tokens on dual NVIDIA A5000 GPUs (24GB each). Experimental results show that SparseAccelerate achieves up to a 1.04x reduction in Time-To-First-Token (TTFT) latency at 32K tokens, while also providing substantial memory savings. These improvements yield practical gains for memory-intensive applications and long-context tasks that were previously infeasible with standard attention. Beyond latency reductions, SparseAccelerate fundamentally shifts the scaling trend, demonstrating the smallest TTFT growth gradient relative to context length among competing methods. Ongoing evaluations on diverse benchmarks confirm its scalability, positioning SparseAccelerate as a critical advancement toward efficient, real-time, and large-context LLM inference on accessible hardware.
☆ Annotations for Exploring Food Tweets From Multiple Aspects
This research builds upon the Latvian Twitter Eater Corpus (LTEC), which is focused on the narrow domain of tweets related to food, drinks, eating and drinking. LTEC has been collected for more than 12 years and reaching almost 3 million tweets with the basic information as well as extended automatically and manually annotated metadata. In this paper we supplement the LTEC with manually annotated subsets of evaluation data for machine translation, named entity recognition, timeline-balanced sentiment analysis, and text-image relation classification. We experiment with each of the data sets using baseline models and highlight future challenges for various modelling approaches.
☆ Query-Efficient Planning with Language Models
Planning in complex environments requires an agent to efficiently query a world model to find a feasible sequence of actions from start to goal. Recent work has shown that Large Language Models (LLMs), with their rich prior knowledge and reasoning capabilities, can potentially help with planning by searching over promising states and adapting to feedback from the world. In this paper, we propose and study two fundamentally competing frameworks that leverage LLMs for query-efficient planning. The first uses LLMs as a heuristic within a search-based planner to select promising nodes to expand and propose promising actions. The second uses LLMs as a generative planner to propose an entire sequence of actions from start to goal, query a world model, and adapt based on feedback. We show that while both approaches improve upon comparable baselines, using an LLM as a generative planner results in significantly fewer interactions. Our key finding is that the LLM as a planner can more rapidly adapt its planning strategies based on immediate feedback than LLM as a heuristic. We present evaluations and ablations on Robotouille and PDDL planning benchmarks and discuss connections to existing theory on query-efficient planning algorithms. Code is available at https://github.com/portal-cornell/llms-for-planning
comment: 11 pages (not including references or appendix); 13 figures (9 main paper, 4 appendix); (v1) preprint
☆ The Computational Limits of State-Space Models and Mamba via the Lens of Circuit Complexity
In this paper, we analyze the computational limitations of Mamba and State-space Models (SSMs) by using the circuit complexity framework. Despite Mamba's stateful design and recent attention as a strong candidate to outperform Transformers, we have demonstrated that both Mamba and SSMs with $\mathrm{poly}(n)$-precision and constant-depth layers reside within the $\mathsf{DLOGTIME}$-uniform $\mathsf{TC}^0$ complexity class. This result indicates Mamba has the same computational capabilities as Transformer theoretically, and it cannot solve problems like arithmetic formula problems, boolean formula value problems, and permutation composition problems if $\mathsf{TC}^0 \neq \mathsf{NC}^1$. Therefore, it challenges the assumption Mamba is more computationally expressive than Transformers. Our contributions include rigorous proofs showing that Selective SSM and Mamba architectures can be simulated by $\mathsf{DLOGTIME}$-uniform $\mathsf{TC}^0$ circuits, and they cannot solve problems outside $\mathsf{TC}^0$.
☆ Hate Speech According to the Law: An Analysis for Effective Detection
The issue of hate speech extends beyond the confines of the online realm. It is a problem with real-life repercussions, prompting most nations to formulate legal frameworks that classify hate speech as a punishable offence. These legal frameworks differ from one country to another, contributing to the big chaos that online platforms have to face when addressing reported instances of hate speech. With the definitions of hate speech falling short in introducing a robust framework, we turn our gaze onto hate speech laws. We consult the opinion of legal experts on a hate speech dataset and we experiment by employing various approaches such as pretrained models both on hate speech and legal data, as well as exploiting two large language models (Qwen2-7B-Instruct and Meta-Llama-3-70B). Due to the time-consuming nature of data acquisition for prosecutable hate speech, we use pseudo-labeling to improve our pretrained models. This study highlights the importance of amplifying research on prosecutable hate speech and provides insights into effective strategies for combating hate speech within the parameters of legal frameworks. Our findings show that legal knowledge in the form of annotations can be useful when classifying prosecutable hate speech, yet more focus should be paid on the differences between the laws.
☆ MMedPO: Aligning Medical Vision-Language Models with Clinical-Aware Multimodal Preference Optimization
The advancement of Large Vision-Language Models (LVLMs) has propelled their application in the medical field. However, Medical LVLMs (Med-LVLMs) encounter factuality challenges due to modality misalignment, where the models prioritize textual knowledge over visual input, leading to hallucinations that contradict information in medical images. Previous attempts to enhance modality alignment in Med-LVLMs through preference optimization have inadequately mitigated clinical relevance in preference data, making these samples easily distinguishable and reducing alignment effectiveness. To address this challenge, we propose MMedPO, a novel multimodal medical preference optimization approach that considers the clinical relevance of preference samples to enhance Med-LVLM alignment. MMedPO curates multimodal preference data by introducing two types of dispreference: (1) plausible hallucinations injected through target Med-LVLMs or GPT-4o to produce medically inaccurate responses, and (2) lesion region neglect achieved through local lesion-noising, disrupting visual understanding of critical areas. We then calculate clinical relevance for each sample based on scores from multiple Med-LLMs and visual tools, and integrate these scores into the preference optimization process as weights, enabling effective alignment. Our experiments demonstrate that MMedPO significantly enhances factual accuracy in Med-LVLMs, achieving substantial improvements over existing preference optimization methods by averaging 14.2% and 51.7% across the Med-VQA and report generation tasks. Our code are available in https://github.com/aiming-lab/MMedPO.
☆ AIDE: Task-Specific Fine Tuning with Attribute Guided Multi-Hop Data Expansion
Fine-tuning large language models (LLMs) for specific tasks requires high-quality, diverse training data relevant to the task. Recent research has leveraged LLMs to synthesize training data, but existing approaches either depend on large seed datasets or struggle to ensure both task relevance and data diversity in the generated outputs. To address these challenges, we propose AIDE, a novel data synthesis framework that uses a multi-hop process to expand 10 seed data points while ensuring diversity and task relevance. AIDE extracts the main topic and key knowledge attributes from the seed data to guide the synthesis process. In each subsequent hop, it extracts the topic and attributes from the newly generated data and continues guided synthesis. This process repeats for a total of K hops. To prevent irrelevant data generation as the hop depth increases, AIDE incorporates a residual connection mechanism and uses self-reflection to improve data quality. Our empirical results demonstrate that fine-tuning Mistral-7B, Llama-3.1-8B and Llama-3.2-3B with AIDE achieves more than 10% accuracy improvements over the base models across 13 tasks from 5 different benchmarks, while outperforming the models fine-tuned with state-of-the-art data synthesis methods like Evol-Instruct, DataTune and Prompt2Model.
comment: 19 pages
☆ Evaluating and Mitigating Social Bias for Large Language Models in Open-ended Settings
Current social bias benchmarks for Large Language Models (LLMs) primarily rely on pre-defined question formats like multiple-choice, limiting their ability to reflect the complexity and open-ended nature of real-world interactions. To address this gap, we extend an existing BBQ dataset introduced by incorporating fill-in-the-blank and short-answer question types, designed to evaluate biases in an open-ended setting. Our finding reveals that LLMs tend to produce responses that are more biased against certain protected attributes, like age and socio-economic status. On the other hand, these biased outputs produced by LLMs can serve as valuable contexts and chains of thought for debiasing. Our debiasing approach combined zero-shot, few-shot, and chain-of-thought could significantly reduce the level of bias to almost 0. We open-source our evaluation and debiasing code hoping to encourage further measurements and mitigation of bias and stereotype in LLMs.
comment: 12 panges
☆ FM2DS: Few-Shot Multimodal Multihop Data Synthesis with Knowledge Distillation for Question Answering CVPR 2025
Multimodal multihop question answering is a complex task that requires reasoning over multiple sources of information, such as images and text, to answer questions. While there has been significant progress in visual question answering, the multihop setting remains unexplored due to the lack of high-quality datasets. Current methods focus on single-hop question answering or a single modality, which makes them unsuitable for real-world scenarios such as analyzing multimodal educational materials, summarizing lengthy academic articles, or interpreting scientific studies that combine charts, images, and text. To address this gap, we propose a novel methodology, introducing the first framework for creating a high-quality dataset that enables training models for multimodal multihop question answering. Our approach consists of a 5-stage pipeline that involves acquiring relevant multimodal documents from Wikipedia, synthetically generating high-level questions and answers, and validating them through rigorous criteria to ensure quality data. We evaluate our methodology by training models on our synthesized dataset and testing on two benchmarks, our results demonstrate that, with an equal sample size, models trained on our synthesized data outperform those trained on human-collected data by 1.9 in exact match (EM) on average. We believe our data synthesis method will serve as a strong foundation for training and evaluating multimodal multihop question answering models.
comment: 20 pages, 11 figures, 10 tables, Submitted to CVPR 2025
☆ Assessing the Impact of Conspiracy Theories Using Large Language Models
Measuring the relative impact of CTs is important for prioritizing responses and allocating resources effectively, especially during crises. However, assessing the actual impact of CTs on the public poses unique challenges. It requires not only the collection of CT-specific knowledge but also diverse information from social, psychological, and cultural dimensions. Recent advancements in large language models (LLMs) suggest their potential utility in this context, not only due to their extensive knowledge from large training corpora but also because they can be harnessed for complex reasoning. In this work, we develop datasets of popular CTs with human-annotated impacts. Borrowing insights from human impact assessment processes, we then design tailored strategies to leverage LLMs for performing human-like CT impact assessments. Through rigorous experiments, we textit{discover that an impact assessment mode using multi-step reasoning to analyze more CT-related evidence critically produces accurate results; and most LLMs demonstrate strong bias, such as assigning higher impacts to CTs presented earlier in the prompt, while generating less accurate impact assessments for emotionally charged and verbose CTs.
☆ Asynchronous LLM Function Calling
Large language models (LLMs) use function calls to interface with external tools and data source. However, the current approach to LLM function calling is inherently synchronous, where each call blocks LLM inference, limiting LLM operation and concurrent function execution. In this work, we propose AsyncLM, a system for asynchronous LLM function calling. AsyncLM improves LLM's operational efficiency by enabling LLMs to generate and execute function calls concurrently. Instead of waiting for each call's completion, AsyncLM introduces an interrupt mechanism to asynchronously notify the LLM in-flight when function calls return. We design an in-context protocol for function calls and interrupts, provide fine-tuning strategy to adapt LLMs to the interrupt semantics, and implement these mechanisms efficiently on LLM inference process. We demonstrate that AsyncLM can reduce end-to-end task completion latency from 1.6x-5.4x compared to synchronous function calling on a set of benchmark tasks in the Berkeley function calling leaderboard (BFCL). Furthermore, we discuss how interrupt mechanisms can be extended to enable novel human-LLM or LLM-LLM interactions.
☆ AutoReason: Automatic Few-Shot Reasoning Decomposition
Chain of Thought (CoT) was introduced in recent research as a method for improving step-by-step reasoning in Large Language Models. However, CoT has limited applications such as its need for hand-crafted few-shot exemplar prompts and no capability to adjust itself to different queries. In this work, we propose a system to automatically generate rationales using CoT. Our method improves multi-step implicit reasoning capabilities by decomposing the implicit query into several explicit questions. This provides interpretability for the model, improving reasoning in weaker LLMs. We test our approach with two Q\&A datasets: StrategyQA and HotpotQA. We show an increase in accuracy with both, especially on StrategyQA. To facilitate further research in this field, the complete source code for this study has been made publicly available on GitHub: https://github.com/miralab-ai/autoreason.
☆ Effective Text Adaptation for LLM-based ASR through Soft Prompt Fine-Tuning
The advent of Large Language Models (LLM) has reformed the Automatic Speech Recognition (ASR). Prompting LLM with audio embeddings to generate transcriptions becomes the new state-of-the-art ASR. Despite LLMs being trained with an extensive amount of text corpora, high-quality domain-specific text data can still significantly enhance ASR performance on domain adaptation tasks. Although LLM-based ASR can naturally incorporate more text corpora by fine-tuning the LLM decoder, fine-tuning such ASR on text-only data without paired prompts may diminish the effectiveness of domain-specific knowledge. To mitigate this issue, we propose a two-step soft prompt fine-tuning strategy that enhances domain-specific text adaptation. Experimental results show that text adaptation with our proposed method achieved a relative up to 9% Word Error Rate (WER) reduction and up to 18% Entity Error Rate (EER) reduction on the target domain compared to the baseline ASR. Combining this with domain-specific Language Model (LM) fusion can further improve the EER by a relative 2-5%
comment: accepted as SLT 2024 proceeding
☆ Analysing Public Transport User Sentiment on Low Resource Multilingual Data
Public transport systems in many Sub-Saharan countries often receive less attention compared to other sectors, underscoring the need for innovative solutions to improve the Quality of Service (QoS) and overall user experience. This study explored commuter opinion mining to understand sentiments toward existing public transport systems in Kenya, Tanzania, and South Africa. We used a qualitative research design, analysing data from X (formerly Twitter) to assess sentiments across rail, mini-bus taxis, and buses. By leveraging Multilingual Opinion Mining techniques, we addressed the linguistic diversity and code-switching present in our dataset, thus demonstrating the application of Natural Language Processing (NLP) in extracting insights from under-resourced languages. We employed PLMs such as AfriBERTa, AfroXLMR, AfroLM, and PuoBERTa to conduct the sentiment analysis. The results revealed predominantly negative sentiments in South Africa and Kenya, while the Tanzanian dataset showed mainly positive sentiments due to the advertising nature of the tweets. Furthermore, feature extraction using the Word2Vec model and K-Means clustering illuminated semantic relationships and primary themes found within the different datasets. By prioritising the analysis of user experiences and sentiments, this research paves the way for developing more responsive, user-centered public transport systems in Sub-Saharan countries, contributing to the broader goal of improving urban mobility and sustainability.
☆ When Every Token Counts: Optimal Segmentation for Low-Resource Language Models COLING 2025
Traditional greedy tokenization methods have been a critical step in Natural Language Processing (NLP), influencing how text is converted into tokens and directly impacting model performance. While subword tokenizers like Byte-Pair Encoding (BPE) are widely used, questions remain about their optimality across model scales and languages. In this work, we demonstrate through extensive experiments that an optimal BPE configuration significantly reduces token count compared to greedy segmentation, yielding improvements in token-saving percentages and performance benefits, particularly for smaller models. We evaluate tokenization performance across various intrinsic and extrinsic tasks, including generation and classification. Our findings suggest that compression-optimized tokenization strategies could provide substantial advantages for multilingual and low-resource language applications, highlighting a promising direction for further research and inclusive NLP.
comment: LoResLM @ COLING 2025
LLMs for Generalizable Language-Conditioned Policy Learning under Minimal Data Requirements
To develop autonomous agents capable of executing complex, multi-step decision-making tasks as specified by humans in natural language, existing reinforcement learning approaches typically require expensive labeled datasets or access to real-time experimentation. Moreover, conventional methods often face difficulties in generalizing to unseen goals and states, thereby limiting their practical applicability. This paper presents TEDUO, a novel training pipeline for offline language-conditioned policy learning. TEDUO operates on easy-to-obtain, unlabeled datasets and is suited for the so-called in-the-wild evaluation, wherein the agent encounters previously unseen goals and states. To address the challenges posed by such data and evaluation settings, our method leverages the prior knowledge and instruction-following capabilities of large language models (LLMs) to enhance the fidelity of pre-collected offline data and enable flexible generalization to new goals and states. Empirical results demonstrate that the dual role of LLMs in our framework-as data enhancers and generalizers-facilitates both effective and data-efficient learning of generalizable language-conditioned policies.
☆ Real-Time Performance Optimization of Travel Reservation Systems Using AI and Microservices
The rapid growth of the travel industry has increased the need for real-time optimization in reservation systems that could take care of huge data and transaction volumes. This study proposes a hybrid framework that ut folds an Artificial Intelligence and a Microservices approach for the performance optimization of the system. The AI algorithms forecast demand patterns, optimize the allocation of resources, and enhance decision-making driven by Microservices architecture, hence decentralizing system components for scalability, fault tolerance, and reduced downtime. The model provided focuses on major problems associated with the travel reservation systems such as latency of systems, load balancing and data consistency. It endows the systems with predictive models based on AI improved ability to forecast user demands. Microservices would also take care of different scales during uneven traffic patterns. Hence, both aspects ensure better handling of peak loads and spikes while minimizing delays and ensuring high service quality. A comparison was made between traditional reservation models, which are monolithic and the new model of AI-Microservices. Comparatively, the analysis results state that there is a drastic improvement in processing times where the system uptime and resource utilization proved the capability of AI and the microservices in transforming the travel industry in terms of reservation. This research work focused on AI and Microservices towards real-time optimization, providing critical insight into how to move forward with practical recommendations for upgrading travel reservation systems with this technology.
comment: 19 pages, 12 figures
♻ ☆ LLM Pruning and Distillation in Practice: The Minitron Approach
We present a comprehensive report on compressing the Llama 3.1 8B and Mistral NeMo 12B models to 4B and 8B parameters, respectively, using pruning and distillation. We explore two distinct pruning strategies: (1) depth pruning and (2) joint hidden/attention/MLP (width) pruning, and evaluate the results on common benchmarks from the LM Evaluation Harness. The models are then aligned with NeMo Aligner and tested in instruct-tuned versions. This approach produces a compelling 4B model from Llama 3.1 8B and a state-of-the-art Mistral-NeMo-Minitron-8B (MN-Minitron-8B for brevity) model from Mistral NeMo 12B. We found that with no access to the original data, it is beneficial to slightly fine-tune teacher models on the distillation dataset. We open-source our base model weights on Hugging Face with a permissive license.
comment: v4: Update author order
♻ ☆ The broader spectrum of in-context learning
The ability of language models to learn a task from a few examples in context has generated substantial interest. Here, we provide a perspective that situates this type of supervised few-shot learning within a much broader spectrum of meta-learned in-context learning. Indeed, we suggest that any distribution of sequences in which context non-trivially decreases loss on subsequent predictions can be interpreted as eliciting a kind of in-context learning. We suggest that this perspective helps to unify the broad set of in-context abilities that language models exhibit $\unicode{x2014}$ such as adapting to tasks from instructions or role play, or extrapolating time series. This perspective also sheds light on potential roots of in-context learning in lower-level processing of linguistic dependencies (e.g. coreference or parallel structures). Finally, taking this perspective highlights the importance of generalization, which we suggest can be studied along several dimensions: not only the ability to learn something novel, but also flexibility in learning from different presentations, and in applying what is learned. We discuss broader connections to past literature in meta-learning and goal-conditioned agents, and other perspectives on learning and adaptation. We close by suggesting that research on in-context learning should consider this broader spectrum of in-context capabilities and types of generalization.
♻ ☆ How Many Languages Make Good Multilingual Instruction Tuning? A Case Study on BLOOM COLING 2025
Instruction tuning a large language model with multiple languages can prepare it for multilingual downstream tasks. Nonetheless, it is yet to be determined whether having a handful of languages is sufficient, or whether the benefits increase with the inclusion of more. By fine-tuning large multilingual models on 1 to 52 languages, we present a case study on BLOOM to understand three pertinent factors affecting performance: the number of languages, language exposure, and similarity between training and test languages. Overall we found that 1) expanding language coverage in multilingual instruction tuning proves to be beneficial; 2) accuracy often significantly boots if the test language appears in the instruction mixture; 3) languages' genetic features correlate with cross-lingual transfer more than merely the number of language but different languages benefit to various degrees.
comment: COLING 2025
♻ ☆ How transformers learn structured data: insights from hierarchical filtering
Understanding the learning process and the embedded computation in transformers is becoming a central goal for the development of interpretable AI. In the present study, we introduce a hierarchical filtering procedure for generative models of sequences on trees, allowing us to hand-tune the range of positional correlations in the data. Leveraging this controlled setting, we provide evidence that vanilla encoder-only transformers can approximate the exact inference algorithm when trained on root classification and masked language modeling tasks, and study how this computation is discovered and implemented. We find that correlations at larger distances, corresponding to increasing layers of the hierarchy, are sequentially included by the network during training. Moreover, by comparing attention maps from models trained with varying degrees of filtering and by probing the different encoder levels, we find clear evidence of a reconstruction of correlations on successive length scales corresponding to the various levels of the hierarchy, which we relate to a plausible implementation of the exact inference algorithm within the same architecture.
comment: 21 pages, 15 figures
♻ ☆ Can tweets predict article retractions? A comparison between human and LLM labelling
Quickly detecting problematic research articles is crucial to safeguarding the integrity of scientific research. This study explores whether Twitter mentions of retracted articles can signal potential problems with the articles prior to their retraction, potentially serving as an early warning system for scholars. To investigate this, we analysed a dataset of 4,354 Twitter mentions associated with 504 retracted articles. The effectiveness of Twitter mentions in predicting article retractions was evaluated by both manual and Large Language Model (LLM) labelling. Manual labelling results indicated that 25.7% of tweets signalled problems before retraction. Using the manual labelling results as the baseline, we found that LLMs (GPT-4o-mini, Gemini 1.5 Flash, and Claude-3.5-Haiku) outperformed lexicon-based sentiment analysis tools (e.g., TextBlob) in detecting potential problems, suggesting that automatic detection of problematic articles from social media using LLMs is technically feasible. Nevertheless, since only a small proportion of retracted articles (11.1%) were criticised on Twitter prior to retraction, such automatic systems would detect only a minority of problematic articles. Overall, this study offers insights into how social media data, coupled with emerging generative AI techniques, can support research integrity.
comment: 19 pages
♻ ☆ How Transformers Solve Propositional Logic Problems: A Mechanistic Analysis
Large language models (LLMs) have shown amazing performance on tasks that require planning and reasoning. Motivated by this, we investigate the internal mechanisms that underpin a network's ability to perform complex logical reasoning. We first construct a synthetic propositional logic problem that serves as a concrete test-bed for network training and evaluation. Crucially, this problem demands nontrivial planning to solve. We perform our study on two fronts. First, we pursue an understanding of precisely how a three-layer transformer, trained from scratch and attains perfect test accuracy, solves this problem. We are able to identify certain "planning" and "reasoning" mechanisms in the network that necessitate cooperation between the attention blocks to implement the desired logic. Second, we study how pretrained LLMs, namely Mistral-7B and Gemma-2-9B, solve this problem. We characterize their reasoning circuits through causal intervention experiments, providing necessity and sufficiency evidence for the circuits. We find evidence suggesting that the two models' latent reasoning strategies are surprisingly similar, and human-like. Overall, our work systemically uncovers novel aspects of small and large transformers, and continues the study of how they plan and reason.
♻ ☆ RARE: Retrieval-Augmented Reasoning Enhancement for Large Language Models
This work introduces RARE (Retrieval-Augmented Reasoning Enhancement), a versatile extension to the mutual reasoning framework (rStar), aimed at enhancing reasoning accuracy and factual integrity across large language models (LLMs) for complex, knowledge-intensive tasks such as commonsense and medical reasoning. RARE incorporates two innovative actions within the Monte Carlo Tree Search (MCTS) framework: A6, which generates search queries based on the initial problem statement, performs information retrieval using those queries, and augments reasoning with the retrieved data to formulate the final answer; and A7, which leverages information retrieval specifically for generated sub-questions and re-answers these sub-questions with the relevant contextual information. Additionally, a Retrieval-Augmented Factuality Scorer is proposed to replace the original discriminator, prioritizing reasoning paths that meet high standards of factuality. Experimental results with LLaMA 3.1 show that RARE enables open-source LLMs to achieve competitive performance with top open-source models like GPT-4 and GPT-4o. This research establishes RARE as a scalable solution for improving LLMs in domains where logical coherence and factual integrity are critical.
comment: 24 pages, 8 figures
♻ ☆ CIC: A Framework for Culturally-Aware Image Captioning IJCAI 2024
Image Captioning generates descriptive sentences from images using Vision-Language Pre-trained models (VLPs) such as BLIP, which has improved greatly. However, current methods lack the generation of detailed descriptive captions for the cultural elements depicted in the images, such as the traditional clothing worn by people from Asian cultural groups. In this paper, we propose a new framework, Culturally-aware Image Captioning (CIC), that generates captions and describes cultural elements extracted from cultural visual elements in images representing cultures. Inspired by methods combining visual modality and Large Language Models (LLMs) through appropriate prompts, our framework (1) generates questions based on cultural categories from images, (2) extracts cultural visual elements from Visual Question Answering (VQA) using generated questions, and (3) generates culturally-aware captions using LLMs with the prompts. Our human evaluation conducted on 45 participants from 4 different cultural groups with a high understanding of the corresponding culture shows that our proposed framework generates more culturally descriptive captions when compared to the image captioning baseline based on VLPs. Resources can be found at https://shane3606.github.io/cic..
comment: Accepted in IJCAI 2024
♻ ☆ Improving the Efficiency of Visually Augmented Language Models COLING 2025
Despite the impressive performance of autoregressive Language Models (LM) it has been shown that due to reporting bias, LMs lack visual knowledge, i.e. they do not know much about the visual world and its properties. To augment LMs with visual knowledge, existing solutions often rely on explicit images, requiring time-consuming retrieval or image generation systems. This paper shows that explicit images are not necessary to visually augment an LM. Instead, we use visually-grounded text representations obtained from the well-known CLIP multimodal system. For a fair comparison, we modify VALM, a visually-augmented LM which uses image retrieval and representation, to work directly with visually-grounded text representations. We name this new model BLIND-VALM. We show that BLIND-VALM performs on par with VALM for Visual Language Understanding (VLU), Natural Language Understanding (NLU) and Language Modeling tasks, despite being significantly more efficient and simpler. We also show that scaling up our model within the compute budget of VALM, either increasing the model or pre-training corpus size, we outperform VALM for all the evaluation tasks.
comment: COLING 2025
♻ ☆ Large Language Models and Games: A Survey and Roadmap
Recent years have seen an explosive increase in research on large language models (LLMs), and accompanying public engagement on the topic. While starting as a niche area within natural language processing, LLMs have shown remarkable potential across a broad range of applications and domains, including games. This paper surveys the current state of the art across the various applications of LLMs in and for games, and identifies the different roles LLMs can take within a game. Importantly, we discuss underexplored areas and promising directions for future uses of LLMs in games and we reconcile the potential and limitations of LLMs within the games domain. As the first comprehensive survey and roadmap at the intersection of LLMs and games, we are hopeful that this paper will serve as the basis for groundbreaking research and innovation in this exciting new field.
comment: Accepted for publication at the IEEE Transactions on Games (19 pages, 6 figures)
♻ ☆ A Survey on Multilingual Large Language Models: Corpora, Alignment, and Bias
Based on the foundation of Large Language Models (LLMs), Multilingual LLMs (MLLMs) have been developed to address the challenges faced in multilingual natural language processing, hoping to achieve knowledge transfer from high-resource languages to low-resource languages. However, significant limitations and challenges still exist, such as language imbalance, multilingual alignment, and inherent bias. In this paper, we aim to provide a comprehensive analysis of MLLMs, delving deeply into discussions surrounding these critical issues. First of all, we start by presenting an overview of MLLMs, covering their evolutions, key techniques, and multilingual capacities. Secondly, we explore the multilingual training corpora of MLLMs and the multilingual datasets oriented for downstream tasks that are crucial to enhance the cross-lingual capability of MLLMs. Thirdly, we survey the state-of-the-art studies of multilingual representations and investigate whether the current MLLMs can learn a universal language representation. Fourthly, we discuss bias on MLLMs, including its categories, evaluation metrics, and debiasing techniques. Finally, we discuss existing challenges and point out promising research directions of MLLMs.
comment: The article has been accepted by Frontiers of Computer Science (FCS), with the DOI: {10.1007/s11704-024-40579-4}
♻ ☆ Imitating Language via Scalable Inverse Reinforcement Learning NeurIPS 2024
The majority of language model training builds on imitation learning. It covers pretraining, supervised fine-tuning, and affects the starting conditions for reinforcement learning from human feedback (RLHF). The simplicity and scalability of maximum likelihood estimation (MLE) for next token prediction led to its role as predominant paradigm. However, the broader field of imitation learning can more effectively utilize the sequential structure underlying autoregressive generation. We focus on investigating the inverse reinforcement learning (IRL) perspective to imitation, extracting rewards and directly optimizing sequences instead of individual token likelihoods and evaluate its benefits for fine-tuning large language models. We provide a new angle, reformulating inverse soft-Q-learning as a temporal difference regularized extension of MLE. This creates a principled connection between MLE and IRL and allows trading off added complexity with increased performance and diversity of generations in the supervised fine-tuning (SFT) setting. We find clear advantages for IRL-based imitation, in particular for retaining diversity while maximizing task performance, rendering IRL a strong alternative on fixed SFT datasets even without online data generation. Our analysis of IRL-extracted reward functions further indicates benefits for more robust reward functions via tighter integration of supervised and preference-based LLM post-training.
comment: Published at NeurIPS 2024
♻ ☆ Jailbreak Attacks and Defenses against Multimodal Generative Models: A Survey
The rapid evolution of multimodal foundation models has led to significant advancements in cross-modal understanding and generation across diverse modalities, including text, images, audio, and video. However, these models remain susceptible to jailbreak attacks, which can bypass built-in safety mechanisms and induce the production of potentially harmful content. Consequently, understanding the methods of jailbreak attacks and existing defense mechanisms is essential to ensure the safe deployment of multimodal generative models in real-world scenarios, particularly in security-sensitive applications. To provide comprehensive insight into this topic, this survey reviews jailbreak and defense in multimodal generative models. First, given the generalized lifecycle of multimodal jailbreak, we systematically explore attacks and corresponding defense strategies across four levels: input, encoder, generator, and output. Based on this analysis, we present a detailed taxonomy of attack methods, defense mechanisms, and evaluation frameworks specific to multimodal generative models. Additionally, we cover a wide range of input-output configurations, including modalities such as Any-to-Text, Any-to-Vision, and Any-to-Any within generative systems. Finally, we highlight current research challenges and propose potential directions for future research. The open-source repository corresponding to this work can be found at https://github.com/liuxuannan/Awesome-Multimodal-Jailbreak.
comment: ongoing work
♻ ☆ Speech Robust Bench: A Robustness Benchmark For Speech Recognition NeurIPS
As Automatic Speech Recognition (ASR) models become ever more pervasive, it is important to ensure that they make reliable predictions under corruptions present in the physical and digital world. We propose Speech Robust Bench (SRB), a comprehensive benchmark for evaluating the robustness of ASR models to diverse corruptions. SRB is composed of 114 input perturbations which simulate an heterogeneous range of corruptions that ASR models may encounter when deployed in the wild. We use SRB to evaluate the robustness of several state-of-the-art ASR models and observe that model size and certain modeling choices such as the use of discrete representations, or self-training appear to be conducive to robustness. We extend this analysis to measure the robustness of ASR models on data from various demographic subgroups, namely English and Spanish speakers, and males and females. Our results revealed noticeable disparities in the model's robustness across subgroups. We believe that SRB will significantly facilitate future research towards robust ASR models, by making it easier to conduct comprehensive and comparable robustness evaluations.
comment: submitted to NeurIPS datasets and benchmark track 2025
♻ ☆ TartuNLP at EvaLatin 2024: Emotion Polarity Detection
This paper presents the TartuNLP team submission to EvaLatin 2024 shared task of the emotion polarity detection for historical Latin texts. Our system relies on two distinct approaches to annotating training data for supervised learning: 1) creating heuristics-based labels by adopting the polarity lexicon provided by the organizers and 2) generating labels with GPT4. We employed parameter efficient fine-tuning using the adapters framework and experimented with both monolingual and cross-lingual knowledge transfer for training language and task adapters. Our submission with the LLM-generated labels achieved the overall first place in the emotion polarity detection task. Our results show that LLM-based annotations show promising results on texts in Latin.
comment: Added Acknowledgments section
♻ ☆ TartuNLP @ SIGTYP 2024 Shared Task: Adapting XLM-RoBERTa for Ancient and Historical Languages
We present our submission to the unconstrained subtask of the SIGTYP 2024 Shared Task on Word Embedding Evaluation for Ancient and Historical Languages for morphological annotation, POS-tagging, lemmatization, character- and word-level gap-filling. We developed a simple, uniform, and computationally lightweight approach based on the adapters framework using parameter-efficient fine-tuning. We applied the same adapter-based approach uniformly to all tasks and 16 languages by fine-tuning stacked language- and task-specific adapters. Our submission obtained an overall second place out of three submissions, with the first place in word-level gap-filling. Our results show the feasibility of adapting language models pre-trained on modern languages to historical and ancient languages via adapter training.
comment: 11 pages, 3 figures, added Acknowledgments section
♻ ☆ On the Universal Truthfulness Hyperplane Inside LLMs
While large language models (LLMs) have demonstrated remarkable abilities across various fields, hallucination remains a significant challenge. Recent studies have explored hallucinations through the lens of internal representations, proposing mechanisms to decipher LLMs' adherence to facts. However, these approaches often fail to generalize to out-of-distribution data, leading to concerns about whether internal representation patterns reflect fundamental factual awareness, or only overfit spurious correlations on the specific datasets. In this work, we investigate whether a universal truthfulness hyperplane that distinguishes the model's factually correct and incorrect outputs exists within the model. To this end, we scale up the number of training datasets and conduct an extensive evaluation -- we train the truthfulness hyperplane on a diverse collection of over 40 datasets and examine its cross-task, cross-domain, and in-domain generalization. Our results indicate that increasing the diversity of the training datasets significantly enhances the performance in all scenarios, while the volume of data samples plays a less critical role. This finding supports the optimistic hypothesis that a universal truthfulness hyperplane may indeed exist within the model, offering promising directions for future research.
♻ ☆ LOLA -- An Open-Source Massively Multilingual Large Language Model
This paper presents LOLA, a massively multilingual large language model trained on more than 160 languages using a sparse Mixture-of-Experts Transformer architecture. Our architectural and implementation choices address the challenge of harnessing linguistic diversity while maintaining efficiency and avoiding the common pitfalls of multilinguality. Our analysis of the evaluation results shows competitive performance in natural language generation and understanding tasks. Additionally, we demonstrate how the learned expert-routing mechanism exploits implicit phylogenetic linguistic patterns to potentially alleviate the curse of multilinguality. We provide an in-depth look at the training process, an analysis of the datasets, and a balanced exploration of the model's strengths and limitations. As an open-source model, LOLA promotes reproducibility and serves as a robust foundation for future research. Our findings enable the development of compute-efficient multilingual models with strong, scalable performance across languages.
♻ ☆ A Topic-level Self-Correctional Approach to Mitigate Hallucinations in MLLMs
Aligning the behaviors of Multimodal Large Language Models (MLLMs) with human preferences is crucial for developing robust and trustworthy AI systems. While recent attempts have employed human experts or powerful auxiliary AI systems to provide more accurate preference feedback, such as determining the preferable responses from MLLMs or directly rewriting hallucination-free responses, extensive resource overhead compromise the scalability of the feedback collection. In this work, we introduce Topic-level Preference Overwriting (TPO), a self-correctional approach that guide the model itself to mitigate its own hallucination at the topic level. Through a deconfounded strategy that replaces each topic within the response with the best or worst alternatives generated by the model itself, TPO creates more contrasting pairwise preference feedback, enhancing the feedback quality without human or proprietary model intervention. Notably, the experimental results demonstrate proposed TPO achieves state-of-the-art performance in trustworthiness, significantly reducing the object hallucinations by 92% and overall hallucinations by 38%. Code, model and dataset are available now.
♻ ☆ Creativity in AI: Progresses and Challenges
Creativity is the ability to produce novel, useful, and surprising ideas, and has been widely studied as a crucial aspect of human cognition. Machine creativity on the other hand has been a long-standing challenge. With the rise of advanced generative AI, there has been renewed interest and debate regarding AI's creative capabilities. Therefore, it is imperative to revisit the state of creativity in AI and identify key progresses and remaining challenges. In this work, we survey leading works studying the creative capabilities of AI systems, focusing on creative problem-solving, linguistic, artistic, and scientific creativity. Our review suggests that while the latest AI models are largely capable of producing linguistically and artistically creative outputs such as poems, images, and musical pieces, they struggle with tasks that require creative problem-solving, abstract thinking and compositionality and their generations suffer from a lack of diversity, originality, long-range incoherence and hallucinations. We also discuss key questions concerning copyright and authorship issues with generative models. Furthermore, we highlight the need for a comprehensive evaluation of creativity that is process-driven and considers several dimensions of creativity. Finally, we propose future research directions to improve the creativity of AI outputs, drawing inspiration from cognitive science and psychology.
comment: minor updates to content + figures
♻ ☆ Lying Blindly: Bypassing ChatGPT's Safeguards to Generate Hard-to-Detect Disinformation Claims
As Large Language Models become more proficient, their misuse in coordinated disinformation campaigns is a growing concern. This study explores the capability of ChatGPT with GPT-3.5 to generate short-form disinformation claims about the war in Ukraine, both in general and on a specific event, which is beyond the GPT-3.5 knowledge cutoff. Unlike prior work, we do not provide the model with human-written disinformation narratives by including them in the prompt. Thus the generated short claims are hallucinations based on prior world knowledge and inference from the minimal prompt. With a straightforward prompting technique, we are able to bypass model safeguards and generate numerous short claims. We compare those against human-authored false claims on the war in Ukraine from ClaimReview, specifically with respect to differences in their linguistic properties. We also evaluate whether AI authorship can be differentiated by human readers or state-of-the-art authorship detection tools. Thus, we demonstrate that ChatGPT can produce realistic, target-specific disinformation claims, even on a specific post-cutoff event, and that they cannot be reliably distinguished by humans or existing automated tools.
♻ ☆ Improved GUI Grounding via Iterative Narrowing
Graphical User Interface (GUI) grounding plays a crucial role in enhancing the capabilities of Vision-Language Model (VLM) agents. While general VLMs, such as GPT-4V, demonstrate strong performance across various tasks, their proficiency in GUI grounding remains suboptimal. Recent studies have focused on fine-tuning these models specifically for one-shot GUI grounding, yielding significant improvements over baseline performance. We introduce a visual prompting framework that employs an iterative narrowing mechanism to further improve the performance of both general and fine-tuned models in GUI grounding. For evaluation, we tested our method on a comprehensive benchmark comprising various UI platforms and provided the code to reproduce our results.
comment: Code available at https://github.com/ant-8/GUI-Grounding-via-Iterative-Narrowing
♻ ☆ Questioning the Survey Responses of Large Language Models NeurIPS 2024
Surveys have recently gained popularity as a tool to study large language models. By comparing survey responses of models to those of human reference populations, researchers aim to infer the demographics, political opinions, or values best represented by current language models. In this work, we critically examine this methodology on the basis of the well-established American Community Survey by the U.S. Census Bureau. Evaluating 43 different language models using de-facto standard prompting methodologies, we establish two dominant patterns. First, models' responses are governed by ordering and labeling biases, for example, towards survey responses labeled with the letter "A". Second, when adjusting for these systematic biases through randomized answer ordering, models across the board trend towards uniformly random survey responses, irrespective of model size or pre-training data. As a result, in contrast to conjectures from prior work, survey-derived alignment measures often permit a simple explanation: models consistently appear to better represent subgroups whose aggregate statistics are closest to uniform for any survey under consideration.
comment: NeurIPS 2024
♻ ☆ Large Language Model Benchmarks in Medical Tasks
With the increasing application of large language models (LLMs) in the medical domain, evaluating these models' performance using benchmark datasets has become crucial. This paper presents a comprehensive survey of various benchmark datasets employed in medical LLM tasks. These datasets span multiple modalities including text, image, and multimodal benchmarks, focusing on different aspects of medical knowledge such as electronic health records (EHRs), doctor-patient dialogues, medical question-answering, and medical image captioning. The survey categorizes the datasets by modality, discussing their significance, data structure, and impact on the development of LLMs for clinical tasks such as diagnosis, report generation, and predictive decision support. Key benchmarks include MIMIC-III, MIMIC-IV, BioASQ, PubMedQA, and CheXpert, which have facilitated advancements in tasks like medical report generation, clinical summarization, and synthetic data generation. The paper summarizes the challenges and opportunities in leveraging these benchmarks for advancing multimodal medical intelligence, emphasizing the need for datasets with a greater degree of language diversity, structured omics data, and innovative approaches to synthesis. This work also provides a foundation for future research in the application of LLMs in medicine, contributing to the evolving field of medical artificial intelligence.
comment: 25 pages, 5 tables
♻ ☆ Eliminating Biased Length Reliance of Direct Preference Optimization via Down-Sampled KL Divergence EMNLP 2024
Direct Preference Optimization (DPO) has emerged as a prominent algorithm for the direct and robust alignment of Large Language Models (LLMs) with human preferences, offering a more straightforward alternative to the complex Reinforcement Learning from Human Feedback (RLHF). Despite its promising efficacy, DPO faces a notable drawback: "verbosity", a common over-optimization phenomenon also observed in RLHF. While previous studies mainly attributed verbosity to biased labels within the data, we propose that the issue also stems from an inherent algorithmic length reliance in DPO. Specifically, we suggest that the discrepancy between sequence-level Kullback-Leibler (KL) divergences between chosen and rejected sequences, used in DPO, results in overestimated or underestimated rewards due to varying token lengths. Empirically, we utilize datasets with different label lengths to demonstrate the presence of biased rewards. We then introduce an effective downsampling approach, named SamPO, to eliminate potential length reliance. Our experimental evaluations, conducted across three LLMs of varying scales and a diverse array of conditional and open-ended benchmarks, highlight the efficacy of SamPO in mitigating verbosity, achieving improvements of 5% to 12% over DPO through debaised rewards. Our codes can be accessed at: https://github.com/LuJunru/SamPO/.
comment: EMNLP 2024 Main, Final Version
♻ ☆ Assessing the potential of LLM-assisted annotation for corpus-based pragmatics and discourse analysis: The case of apology
Certain forms of linguistic annotation, like part of speech and semantic tagging, can be automated with high accuracy. However, manual annotation is still necessary for complex pragmatic and discursive features that lack a direct mapping to lexical forms. This manual process is time-consuming and error-prone, limiting the scalability of function-to-form approaches in corpus linguistics. To address this, our study explores the possibility of using large language models (LLMs) to automate pragma-discursive corpus annotation. We compare GPT-3.5 (the model behind the free-to-use version of ChatGPT), GPT-4 (the model underpinning the precise mode of Bing chatbot), and a human coder in annotating apology components in English based on the local grammar framework. We find that GPT-4 outperformed GPT-3.5, with accuracy approaching that of a human coder. These results suggest that LLMs can be successfully deployed to aid pragma-discursive corpus annotation, making the process more efficient, scalable and accessible.
comment: 24 pages, 2 figures, 3 tablels
♻ ☆ FIPO: Free-form Instruction-oriented Prompt Optimization with Preference Dataset and Modular Fine-tuning Schema COLING 2025
When the quality of naive prompts is carefully optimized by human experts, the task performance of large language models (LLMs) can be significantly improved. However, expert-based prompt optimizations are expensive. Herein, some works have proposed Automatic Prompt Optimization (APO), to optimize naive prompts according to task outputs of given in-box testing models, with the help of advanced LLMs (e.g., GPT-4) in an ad-hoc way. Although effective, existing schemes suffer from poor generalization ability and privacy risk. To this end, we collect the first large-scale Prompt Optimization Preference dataset (POP), fine-tune offline local LLM-based optimizers, then fairly test with various downstream models. Our method allows accurate optimization of the core task instruction part within the naive prompt in a model-agnostic manner, and thus is named Free-from Instruction-oriented Prompt Optimization (FIPO). In specific, FIPO uses a modular APO template that dynamically integrate the naive task instruction, optional instruction responses, and optional ground truth to produce finely optimized prompts. The POP dataset is meticulously constructed using advanced LLMs, undergoing rigorous cross-validation by human experts and analytical models. Leveraging insights from the data with Tulu2 models and diverse fine-tuning strategies, we validate the efficacy of FIPO framework across five public benchmarks and six testing models. Check codes and data here: https://github.com/LuJunru/FIPO_Project.
comment: COLING 2025, Final Version
♻ ☆ Split and Merge: Aligning Position Biases in LLM-based Evaluators EMNLP 2024
Large language models (LLMs) have shown promise as automated evaluators for assessing the quality of answers generated by AI systems. However, these LLM-based evaluators exhibit position bias, or inconsistency, when used to evaluate candidate answers in pairwise comparisons, favoring either the first or second answer regardless of content. To address this limitation, we propose PORTIA, an alignment-based system designed to mimic human comparison strategies to calibrate position bias in a lightweight yet effective manner. Specifically, PORTIA splits the answers into multiple segments, aligns similar content across candidate answers, and then merges them back into a single prompt for evaluation by LLMs. We conducted extensive experiments with six diverse LLMs to evaluate 11,520 answer pairs. Our results show that PORTIA markedly enhances the consistency rates for all the models and comparison forms tested, achieving an average relative improvement of 47.46%. Remarkably, PORTIA enables less advanced GPT models to achieve 88% agreement with the state-of-the-art GPT-4 model at just 10% of the cost. Furthermore, it rectifies around 80% of the position bias instances within the GPT-4 model, elevating its consistency rate up to 98%. Subsequent human evaluations indicate that the PORTIA-enhanced GPT-3.5 model can even surpass the standalone GPT-4 in terms of alignment with human evaluators. These findings highlight PORTIA's ability to correct position bias, improve LLM consistency, and boost performance while keeping cost-efficiency. This represents a valuable step toward a more reliable and scalable use of LLMs for automated evaluations across diverse applications.
comment: Accepted by EMNLP 2024. Please cite the conference version of this paper, e.g., "Zongjie Li, Chaozheng Wang, Pingchuan Ma, Daoyuan Wu, Shuai Wang, Cuiyun Gao, and Yang Liu. 2024. Split and Merge: Aligning Position Biases in LLM-based Evaluators. (EMNLP 2024)"
♻ ☆ Rethinking Data Selection at Scale: Random Selection is Almost All You Need
Supervised fine-tuning (SFT) is crucial for aligning Large Language Models (LLMs) with human instructions. The primary goal during SFT is to select a small yet representative subset of training data from the larger pool, such that fine-tuning with this subset achieves results comparable to or even exceeding those obtained using the entire dataset. However, most existing data selection techniques are designed for small-scale data pools, which fail to meet the demands of real-world SFT scenarios. In this paper, we replicated several self-scoring methods those that do not rely on external model assistance on two million scale datasets, and found that nearly all methods struggled to significantly outperform random selection when dealing with such large-scale data pools. Moreover, our comparisons suggest that, during SFT, diversity in data selection is more critical than simply focusing on high quality data. We also analyzed the limitations of several current approaches, explaining why they perform poorly on large-scale datasets and why they are unsuitable for such contexts. Finally, we found that filtering data by token length offers a stable and efficient method for improving results. This approach, particularly when training on long text data, proves highly beneficial for relatively weaker base models, such as Llama3.
♻ ☆ EXAONE 3.5: Series of Large Language Models for Real-world Use Cases
This technical report introduces the EXAONE 3.5 instruction-tuned language models, developed and released by LG AI Research. The EXAONE 3.5 language models are offered in three configurations: 32B, 7.8B, and 2.4B. These models feature several standout capabilities: 1) exceptional instruction following capabilities in real-world scenarios, achieving the highest scores across seven benchmarks, 2) outstanding long-context comprehension, attaining the top performance in four benchmarks, and 3) competitive results compared to state-of-the-art open models of similar sizes across nine general benchmarks. The EXAONE 3.5 language models are open to anyone for research purposes and can be downloaded from https://huggingface.co/LGAI-EXAONE. For commercial use, please reach out to the official contact point of LG AI Research: contact_us@lgresearch.ai.
comment: arXiv admin note: text overlap with arXiv:2408.03541
♻ ☆ Enhancing FKG.in: automating Indian food composition analysis
This paper presents a novel approach to compute food composition data for Indian recipes using a knowledge graph for Indian food (FKG.in) and LLMs. The primary focus is to provide a broad overview of an automated food composition analysis workflow and describe its core functionalities: nutrition data aggregation, food composition analysis, and LLM-augmented information resolution. This workflow aims to complement FKG.in and iteratively supplement food composition data from verified knowledge bases. Additionally, this paper highlights the challenges of representing Indian food and accessing food composition data digitally. It also reviews three key sources of food composition data: the Indian Food Composition Tables, the Indian Nutrient Databank, and the Nutritionix API. Furthermore, it briefly outlines how users can interact with the workflow to obtain diet-based health recommendations and detailed food composition information for numerous recipes. We then explore the complex challenges of analyzing Indian recipe information across dimensions such as structure, multilingualism, and uncertainty as well as present our ongoing work on LLM-based solutions to address these issues. The methods proposed in this workshop paper for AI-driven knowledge curation and information resolution are application-agnostic, generalizable, and replicable for any domain.
comment: 15 pages, 5 figures, 30 references, International Conference on Pattern Recognition 2024 - Multimedia Assisted Dietary Management Workshop
♻ ☆ Docling Technical Report
This technical report introduces Docling, an easy to use, self-contained, MIT-licensed open-source package for PDF document conversion. It is powered by state-of-the-art specialized AI models for layout analysis (DocLayNet) and table structure recognition (TableFormer), and runs efficiently on commodity hardware in a small resource budget. The code interface allows for easy extensibility and addition of new features and models.
comment: Docling v1 report
♻ ☆ From Novice to Expert: LLM Agent Policy Optimization via Step-wise Reinforcement Learning
The outstanding capabilities of large language models (LLMs) render them a crucial component in various autonomous agent systems. While traditional methods depend on the inherent knowledge of LLMs without fine-tuning, more recent approaches have shifted toward the reinforcement learning strategy to further enhance agents' ability to solve complex interactive tasks with environments and tools. However, previous approaches are constrained by the sparse reward issue, where existing datasets solely provide a final scalar reward for each multi-step reasoning chain, potentially leading to ineffectiveness and inefficiency in policy learning. In this paper, we introduce StepAgent, which utilizes step-wise reward to optimize the agent's reinforcement learning process. Inheriting the spirit of novice-to-expert theory, we first compare the actions of the expert and the agent to automatically generate intermediate rewards for fine-grained optimization. Additionally, we propose implicit-reward and inverse reinforcement learning techniques to facilitate agent reflection and policy adjustment. Further theoretical analysis demonstrates that the action distribution of the agent can converge toward the expert action distribution over multiple training cycles. Experimental results across various datasets indicate that StepAgent outperforms existing baseline methods.
♻ ☆ psifx -- Psychological and Social Interactions Feature Extraction Package
psifx is a plug-and-play multi-modal feature extraction toolkit, aiming to facilitate and democratize the use of state-of-the-art machine learning techniques for human sciences research. It is motivated by a need (a) to automate and standardize data annotation processes, otherwise involving expensive, lengthy, and inconsistent human labor, such as the transcription or coding of behavior changes from audio and video sources; (b) to develop and distribute open-source community-driven psychology research software; and (c) to enable large-scale access and ease of use to non-expert users. The framework contains an array of tools for tasks, such as speaker diarization, closed-caption transcription and translation from audio, as well as body, hand, and facial pose estimation and gaze tracking from video. The package has been designed with a modular and task-oriented approach, enabling the community to add or update new tools easily. We strongly hope that this package will provide psychologists a simple and practical solution for efficiently a range of audio, linguistic, and visual features from audio and video, thereby creating new opportunities for in-depth study of real-time behavioral phenomena.
♻ ☆ SAM4MLLM: Enhance Multi-Modal Large Language Model for Referring Expression Segmentation
We introduce SAM4MLLM, an innovative approach which integrates the Segment Anything Model (SAM) with Multi-Modal Large Language Models (MLLMs) for pixel-aware tasks. Our method enables MLLMs to learn pixel-level location information without requiring excessive modifications to the existing model architecture or adding specialized tokens. We introduce an inquiry-based approach that can effectively find prompt points for SAM to perform segmentation based on MLLM. It combines detailed visual information with the powerful expressive capabilities of large language models in a unified language-based manner without additional computational overhead in learning. Experimental results on pubic benchmarks demonstrate the effectiveness of our approach.
♻ ☆ Rethinking Data Synthesis: A Teacher Model Training Recipe with Interpretation
Recent advances in large language model (LLM) training have highlighted the need for diverse, high-quality instruction data. Recently, many works are exploring synthetic data generation using LLMs. However, they primarily focus on prompt engineering with standard supervised instruction-finetuned models, which contains a fundamental limitation: these models are optimized for general question-answering/problem-solving rather than data generation. We propose a paradigm shift named \textbf{NOMAD} by investigating how to specifically train models for data generation, demonstrating that this task differs significantly from training a classical LM. We identify two key factors: no-prompt-masked training and proper training set size selection. Our method, NOMAD, shows substantial improvements over baselines, achieving >4\% gains in TriviaQA and >2\% in GSM8K with limited training data. Finally, we offer new insights by interpreting synthetic data through the lenses of "relevance" and "novelty".
♻ ☆ KorMedMCQA: Multi-Choice Question Answering Benchmark for Korean Healthcare Professional Licensing Examinations
We present KorMedMCQA, the first Korean Medical Multiple-Choice Question Answering benchmark, derived from professional healthcare licensing examinations conducted in Korea between 2012 and 2024. The dataset contains 7,469 questions from examinations for doctor, nurse, pharmacist, and dentist, covering a wide range of medical disciplines. We evaluate the performance of 59 large language models, spanning proprietary and open-source models, multilingual and Korean-specialized models, and those fine-tuned for clinical applications. Our results show that applying Chain of Thought (CoT) reasoning can enhance the model performance by up to 4.5% compared to direct answering approaches. We also investigate whether MedQA, one of the most widely used medical benchmarks derived from the U.S. Medical Licensing Examination, can serve as a reliable proxy for evaluating model performance in other regions-in this case, Korea. Our correlation analysis between model scores on KorMedMCQA and MedQA reveals that these two benchmarks align no better than benchmarks from entirely different domains (e.g., MedQA and MMLU-Pro). This finding underscores the substantial linguistic and clinical differences between Korean and U.S. medical contexts, reinforcing the need for region-specific medical QA benchmarks. To support ongoing research in Korean healthcare AI, we publicly release the KorMedMCQA via Huggingface.
♻ ☆ Towards Efficient and Robust VQA-NLE Data Generation with Large Vision-Language Models
Natural Language Explanation (NLE) aims to elucidate the decision-making process by providing detailed, human-friendly explanations in natural language. It helps demystify the decision-making processes of large vision-language models (LVLMs) through the use of language models. While existing methods for creating a Vision Question-Answering with Natural Language Explanation (VQA-NLE) datasets can provide explanations, they heavily rely on human annotations that are time-consuming and costly. In this study, we propose a novel approach that leverages LVLMs to efficiently generate high-quality synthetic VQA-NLE datasets. By evaluating our synthetic data, we showcase how advanced prompting techniques can lead to the production of high-quality VQA-NLE data. Our findings indicate that this proposed method achieves up to 20x faster than human annotation, with only a minimal decrease in qualitative metrics, achieving robust quality that is nearly equivalent to human-annotated data. Furthermore, we show that incorporating visual prompts significantly enhances the relevance of text generation. Our study paves the way for a more efficient and robust automated generation of multi-modal NLE data, offering a promising solution to the problem.
comment: Preprint
♻ ☆ xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token
This paper introduces xRAG, an innovative context compression method tailored for retrieval-augmented generation. xRAG reinterprets document embeddings in dense retrieval--traditionally used solely for retrieval--as features from the retrieval modality. By employing a modality fusion methodology, xRAG seamlessly integrates these embeddings into the language model representation space, effectively eliminating the need for their textual counterparts and achieving an extreme compression rate. In xRAG, the only trainable component is the modality bridge, while both the retriever and the language model remain frozen. This design choice allows for the reuse of offline-constructed document embeddings and preserves the plug-and-play nature of retrieval augmentation. Experimental results demonstrate that xRAG achieves an average improvement of over 10% across six knowledge-intensive tasks, adaptable to various language model backbones, ranging from a dense 7B model to an 8x7B Mixture of Experts configuration. xRAG not only significantly outperforms previous context compression methods but also matches the performance of uncompressed models on several datasets, while reducing overall FLOPs by a factor of 3.53. Our work pioneers new directions in retrieval-augmented generation from the perspective of multimodality fusion, and we hope it lays the foundation for future efficient and scalable retrieval-augmented systems
comment: Neurips 2024
♻ ☆ Exo2EgoDVC: Dense Video Captioning of Egocentric Procedural Activities Using Web Instructional Videos WACV 2025
We propose a novel benchmark for cross-view knowledge transfer of dense video captioning, adapting models from web instructional videos with exocentric views to an egocentric view. While dense video captioning (predicting time segments and their captions) is primarily studied with exocentric videos (e.g., YouCook2), benchmarks with egocentric videos are restricted due to data scarcity. To overcome the limited video availability, transferring knowledge from abundant exocentric web videos is demanded as a practical approach. However, learning the correspondence between exocentric and egocentric views is difficult due to their dynamic view changes. The web videos contain shots showing either full-body or hand regions, while the egocentric view is constantly shifting. This necessitates the in-depth study of cross-view transfer under complex view changes. To this end, we first create a real-life egocentric dataset (EgoYC2) whose captions follow the definition of YouCook2 captions, enabling transfer learning between these datasets with access to their ground-truth. To bridge the view gaps, we propose a view-invariant learning method using adversarial training, which consists of pre-training and fine-tuning stages. Our experiments confirm the effectiveness of overcoming the view change problem and knowledge transfer to egocentric views. Our benchmark pushes the study of cross-view transfer into a new task domain of dense video captioning and envisions methodologies that describe egocentric videos in natural language.
comment: Accepted to WACV 2025. Project page: https://tkhkaeio.github.io/projects/25-egodvc/
♻ ☆ Rethinking Reward Model Evaluation: Are We Barking up the Wrong Tree?
Reward Models (RMs) are crucial for aligning language models with human preferences. Currently, the evaluation of RMs depends on measuring accuracy against a validation set of manually annotated preference data. Although this method is straightforward and widely adopted, the relationship between RM accuracy and downstream policy performance remains under-explored. In this work, we conduct experiments in a synthetic setting to investigate how differences in RM measured by accuracy translate into gaps in optimized policy performance. Our findings reveal that while there is a weak positive correlation between accuracy and downstream performance, policies optimized towards RMs with similar accuracy can exhibit quite different performance. Moreover, we discover that the way of measuring accuracy significantly impacts its ability to predict the final policy performance. Through the lens of the Regressional Goodhart effect, we recognize that accuracy, when used for measuring RM quality, can fail to fully capture the potential RM overoptimization. This underscores the inadequacy of relying solely on accuracy to reflect their impact on policy optimization.
♻ ☆ Pangea: A Fully Open Multilingual Multimodal LLM for 39 Languages
Despite recent advances in multimodal large language models (MLLMs), their development has predominantly focused on English- and western-centric datasets and tasks, leaving most of the world's languages and diverse cultural contexts underrepresented. This paper introduces Pangea, a multilingual multimodal LLM trained on PangeaIns, a diverse 6M instruction dataset spanning 39 languages. PangeaIns features: 1) high-quality English instructions, 2) carefully machine-translated instructions, and 3) culturally relevant multimodal tasks to ensure cross-cultural coverage. To rigorously assess models' capabilities, we introduce PangeaBench, a holistic evaluation suite encompassing 14 datasets covering 47 languages. Results show that Pangea significantly outperforms existing open-source models in multilingual settings and diverse cultural contexts. Ablation studies further reveal the importance of English data proportions, language popularity, and the number of multimodal training samples on overall performance. We fully open-source our data, code, and trained checkpoints, to facilitate the development of inclusive and robust multilingual MLLMs, promoting equity and accessibility across a broader linguistic and cultural spectrum.
comment: 54 pages, 27 figures
♻ ☆ Personalized Language Modeling from Personalized Human Feedback
Personalized large language models (LLMs) are designed to tailor responses to individual user preferences. While Reinforcement Learning from Human Feedback (RLHF) is a commonly used framework for aligning LLMs with human preferences, vanilla RLHF assumes that all human preferences share the same distribution, preventing fine-tuned LLMs from generating personalized content when user preferences are diverse. In this work, we propose Personalized-RLHF (P-RLHF), an efficient framework that utilizes a lightweight user model to capture individual user preferences and jointly learns the user model and the personalized LLM from human feedback. P-RLHF exhibits the following three characteristics: (1) It enables an LLM to generate personalized content and scale efficiently with growing number of users. (2) It handles both explicit user preferences described as textual input and implicit user preferences encoded in the feedback data. (3) It eliminates the need for users to fully articulate their preferences, which are normally needed for prompting LLMs to generate personalized content yet are often impractical to obtain in real-world scenarios. Our experimental results show that personalized LLMs trained using P-RLHF generate responses that are more closely aligned with individual user preferences, outperforming vanilla, non-personalized RLHF and prompting-based personalization approaches across different tasks. We opensource our code at https://github.com/HumainLab/Personalized_RLHF.
♻ ☆ Householder Pseudo-Rotation: A Novel Approach to Activation Editing in LLMs with Direction-Magnitude Perspective EMNLP 2024
Activation Editing, which involves directly editting the internal representations of large language models (LLMs) to alter their behaviors and achieve desired properties, has emerged as a promising area of research. Existing works primarily treat LLMs' activations as points in space and modify them by adding steering vectors. However, this approach is limited in its ability to achieve greater performance improvement while maintaining the necessary consistency of activation magnitudes. To overcome these issues, we propose a novel editing method that views activations in terms of their directions and magnitudes. Our method, named Householder Pseudo-Rotation (HPR), mimics the rotation transformation, thus preserving activation norms and resulting in an improved performance on various safety benchmarks.
comment: EMNLP 2024
♻ ☆ Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation
Retrieval-augmented generation (RAG) has improved large language models (LLMs) by using knowledge retrieval to overcome knowledge deficiencies. However, current RAG methods often fall short of ensuring the depth and completeness of retrieved information, which is necessary for complex reasoning tasks. In this work, we introduce Think-on-Graph 2.0 (ToG-2), a hybrid RAG framework that iteratively retrieves information from both unstructured and structured knowledge sources in a tight-coupling manner. Specifically, ToG-2 leverages knowledge graphs (KGs) to link documents via entities, facilitating deep and knowledge-guided context retrieval. Simultaneously, it utilizes documents as entity contexts to achieve precise and efficient graph retrieval. ToG-2 alternates between graph retrieval and context retrieval to search for in-depth clues relevant to the question, enabling LLMs to generate answers. We conduct a series of well-designed experiments to highlight the following advantages of ToG-2: 1) ToG-2 tightly couples the processes of context retrieval and graph retrieval, deepening context retrieval via the KG while enabling reliable graph retrieval based on contexts; 2) it achieves deep and faithful reasoning in LLMs through an iterative knowledge retrieval process of collaboration between contexts and the KG; and 3) ToG-2 is training-free and plug-and-play compatible with various LLMs. Extensive experiments demonstrate that ToG-2 achieves overall state-of-the-art (SOTA) performance on 6 out of 7 knowledge-intensive datasets with GPT-3.5, and can elevate the performance of smaller models (e.g., LLAMA-2-13B) to the level of GPT-3.5's direct reasoning. The source code is available on https://github.com/IDEA-FinAI/ToG-2.
♻ ☆ Trustful LLMs: Customizing and Grounding Text Generation with Knowledge Bases and Dual Decoders
Although people are impressed by the content generation skills of large language models, the use of LLMs, such as ChatGPT, is limited by the domain grounding of the content. The correctness and groundedness of the generated content need to be based on a verified context, such as results from Retrieval-Augmented Generation (RAG). One important issue when adapting LLMs to a customized domain is that the generated responses are often incomplete, or the additions are not verified and may even be hallucinated. Prior studies on hallucination detection have focused on evaluation metrics, which are not easily adaptable to dynamic domains and can be vulnerable to attacks like jail-breaking. In this work, we propose 1) a post-processing algorithm that leverages knowledge triplets in RAG context to correct hallucinations and 2) a dual-decoder model that fuses RAG context to guide the generation process.
♻ ☆ Opportunities and Challenges of Large Language Models for Low-Resource Languages in Humanities Research
Low-resource languages serve as invaluable repositories of human history, embodying cultural evolution and intellectual diversity. Despite their significance, these languages face critical challenges, including data scarcity and technological limitations, which hinder their comprehensive study and preservation. Recent advancements in large language models (LLMs) offer transformative opportunities for addressing these challenges, enabling innovative methodologies in linguistic, historical, and cultural research. This study systematically evaluates the applications of LLMs in low-resource language research, encompassing linguistic variation, historical documentation, cultural expressions, and literary analysis. By analyzing technical frameworks, current methodologies, and ethical considerations, this paper identifies key challenges such as data accessibility, model adaptability, and cultural sensitivity. Given the cultural, historical, and linguistic richness inherent in low-resource languages, this work emphasizes interdisciplinary collaboration and the development of customized models as promising avenues for advancing research in this domain. By underscoring the potential of integrating artificial intelligence with the humanities to preserve and study humanity's linguistic and cultural heritage, this study fosters global efforts towards safeguarding intellectual diversity.
♻ ☆ Language Control Diffusion: Efficiently Scaling through Space, Time, and Tasks ICLR 2024
Training generalist agents is difficult across several axes, requiring us to deal with high-dimensional inputs (space), long horizons (time), and generalization to novel tasks. Recent advances with architectures have allowed for improved scaling along one or two of these axes, but are still computationally prohibitive to use. In this paper, we propose to address all three axes by leveraging \textbf{L}anguage to \textbf{C}ontrol \textbf{D}iffusion models as a hierarchical planner conditioned on language (LCD). We effectively and efficiently scale diffusion models for planning in extended temporal, state, and task dimensions to tackle long horizon control problems conditioned on natural language instructions, as a step towards generalist agents. Comparing LCD with other state-of-the-art models on the CALVIN language robotics benchmark finds that LCD outperforms other SOTA methods in multi-task success rates, whilst improving inference speed over other comparable diffusion models by 3.3x~15x. We show that LCD can successfully leverage the unique strength of diffusion models to produce coherent long range plans while addressing their weakness in generating low-level details and control.
comment: ICLR 2024, Project and code available at https://github.com/ezhang7423/language-control-diffusion
♻ ☆ Efficacy of Language Model Self-Play in Non-Zero-Sum Games
Game-playing agents like AlphaGo have achieved superhuman performance through self-play, which is theoretically guaranteed to yield optimal policies in competitive games. However, most language tasks are partially or fully cooperative, so it is an open question whether techniques like self-play can effectively be used to improve language models. We empirically investigate this question in a negotiation game setting known as Deal or No Deal (DoND). Crucially, the objective in DoND can be modified to produce a fully cooperative game, a strictly competitive one, or anything in between. We finetune language models in self-play over multiple rounds of filtered behavior cloning in DoND for each of these objectives and evaluate them in self-play and in collaboration with humans. We find that language models improve substantially in self-play, achieving 14-17x higher scores in task reward after finetuning. Further, the trained models generalize to both cooperation and competition with humans, scoring 2.5-6x higher than base models. We view these results as an early promising sign for language model self-play in cooperative settings, despite a lack of theoretical guarantees.
♻ ☆ A Survey on Medical Large Language Models: Technology, Application, Trustworthiness, and Future Directions
With the advent of Large Language Models (LLMs), medical artificial intelligence (AI) has experienced substantial technological progress and paradigm shifts, highlighting the potential of LLMs to streamline healthcare delivery and improve patient outcomes. Considering this rapid technical progress, in this survey, we trace the recent advances of Medical Large Language Models (Med-LLMs), including the background, key findings, and mainstream techniques, especially for the evolution from general-purpose models to medical-specialized applications. Firstly, we delve into the foundational technology of Med-LLMs, indicating how general models can be progressively adapted and refined for the complicated medical tasks. Secondly, the wide-ranging applications of Med-LLMs are investigated across various healthcare domains, as well as an up-to-date review of existing Med-LLMs. The transformative impact of these models on daily medical practice is evident through their ability to assist clinicians, educators, and patients. Recognizing the importance of responsible innovation, we discuss the challenges associated with ensuring fairness, accountability, privacy, and robustness. Ethical considerations, rigorous evaluation methodologies, and the establishment of regulatory frameworks are crucial for building trustworthiness in the real-world system. We emphasize the need for ongoing scrutiny and development to maintain high standards of safety and reliability. Finally, we anticipate possible future trajectories for Med-LLMs, identifying key avenues for prudent expansion. By consolidating these insights, our review aims to provide professionals and researchers with a thorough understanding of the strengths and limitations of Med-LLMs, fostering a balanced and ethical approach to their integration into the healthcare ecosystem.
♻ ☆ Integrative Decoding: Improve Factuality via Implicit Self-consistency
Self-consistency-based approaches, which involve repeatedly sampling multiple outputs and selecting the most consistent one as the final response, prove to be remarkably effective in improving the factual accuracy of large language models. Nonetheless, existing methods usually have strict constraints on the task format, largely limiting their applicability. In this paper, we present Integrative Decoding (ID), to unlock the potential of self-consistency in open-ended generation tasks. ID operates by constructing a set of inputs, each prepended with a previously sampled response, and then processes them concurrently, with the next token being selected by aggregating of all their corresponding predictions at each decoding step. In essence, this simple approach implicitly incorporates self-consistency in the decoding objective. Extensive evaluation shows that ID consistently enhances factuality over a wide range of language models, with substantial improvements on the TruthfulQA (+11.2%), Biographies (+15.4%) and LongFact (+8.5%) benchmarks. The performance gains amplify progressively as the number of sampled responses increases, indicating the potential of ID to scale up with repeated sampling.
♻ ☆ Probing the Decision Boundaries of In-context Learning in Large Language Models NeurIPS 2024
In-context learning is a key paradigm in large language models (LLMs) that enables them to generalize to new tasks and domains by simply prompting these models with a few exemplars without explicit parameter updates. Many attempts have been made to understand in-context learning in LLMs as a function of model scale, pretraining data, and other factors. In this work, we propose a new mechanism to probe and understand in-context learning from the lens of decision boundaries for in-context binary classification. Decision boundaries are straightforward to visualize and provide important information about the qualitative behavior of the inductive biases of standard classifiers. To our surprise, we find that the decision boundaries learned by current LLMs in simple binary classification tasks are often irregular and non-smooth, regardless of linear separability in the underlying task. This paper investigates the factors influencing these decision boundaries and explores methods to enhance their generalizability. We assess various approaches, including training-free and fine-tuning methods for LLMs, the impact of model architecture, and the effectiveness of active prompting techniques for smoothing decision boundaries in a data-efficient manner. Our findings provide a deeper understanding of in-context learning dynamics and offer practical improvements for enhancing robustness and generalizability of in-context learning.
comment: Accepted at NeurIPS 2024, code at https://github.com/siyan-zhao/ICL_decision_boundary
♻ ☆ BPDec: Unveiling the Potential of Masked Language Modeling Decoder in BERT pretraining
BERT (Bidirectional Encoder Representations from Transformers) has revolutionized the field of natural language processing through its exceptional performance on numerous tasks. Yet, the majority of researchers have mainly concentrated on enhancements related to the model structure, such as relative position embedding and more efficient attention mechanisms. Others have delved into pretraining tricks associated with Masked Language Modeling, including whole word masking. DeBERTa introduced an enhanced decoder adapted for BERT's encoder model for pretraining, proving to be highly effective. We argue that the design and research around enhanced masked language modeling decoders have been underappreciated. In this paper, we propose several designs of enhanced decoders and introduce BPDec (BERT Pretraining Decoder), a novel method for modeling training. Typically, a pretrained BERT model is fine-tuned for specific Natural Language Understanding (NLU) tasks. In our approach, we utilize the original BERT model as the encoder, making only changes to the decoder without altering the encoder. This approach does not necessitate extensive modifications to the encoder architecture and can be seamlessly integrated into existing fine-tuning pipelines and services, offering an efficient and effective enhancement strategy. Compared to other methods, while we also incur a moderate training cost for the decoder during the pretraining process, our approach does not introduce additional training costs during the fine-tuning phase. We test multiple enhanced decoder structures after pretraining and evaluate their performance on the GLUE tasks and SQuAD tasks. Our results demonstrate that BPDec, having only undergone subtle refinements to the model structure during pretraining, significantly enhances model performance without escalating the finetuning cost, inference time and serving budget.
♻ ☆ A Video is Worth 10,000 Words: Training and Benchmarking with Diverse Captions for Better Long Video Retrieval WACV 2025
Existing long video retrieval systems are trained and tested in the paragraph-to-video retrieval regime, where every long video is described by a single long paragraph. This neglects the richness and variety of possible valid descriptions of a video, which could range anywhere from moment-by-moment detail to a single phrase summary. To provide a more thorough evaluation of the capabilities of long video retrieval systems, we propose a pipeline that leverages state-of-the-art large language models to carefully generate a diverse set of synthetic captions for long videos. We validate this pipeline's fidelity via rigorous human inspection. We use synthetic captions from this pipeline to perform a benchmark of a representative set of video language models using long video datasets, and show that the models struggle on shorter captions. We show that finetuning on this data can both mitigate these issues (+2.8% R@1 over SOTA on ActivityNet with diverse captions), and even improve performance on standard paragraph-to-video retrieval (+1.0% R@1 on ActivityNet). We also use synthetic data from our pipeline as query expansion in the zero-shot setting (+3.4% R@1 on ActivityNet). We derive insights by analyzing failure cases for retrieval with short captions. For data access and other details, please refer to our project website at https://mgwillia.github.io/10k-words.
comment: 17 pages, 16 tables, 8 figures. To appear at WACV 2025
♻ ☆ Surveying the Effects of Quality, Diversity, and Complexity in Synthetic Data From Large Language Models
Synthetic data generation with Large Language Models is a promising paradigm for augmenting natural data over a nearly infinite range of tasks. Given this variety, direct comparisons among synthetic data generation algorithms are scarce, making it difficult to understand where improvement comes from and what bottlenecks exist. We propose to evaluate algorithms via the makeup of synthetic data generated by each algorithm in terms of data quality, diversity, and complexity. We choose these three characteristics for their significance in open-ended processes and the impact each has on the capabilities of downstream models. We find quality to be essential for in-distribution model generalization, diversity to be essential for out-of-distribution generalization, and complexity to be beneficial for both. Further, we emphasize the existence of Quality-Diversity trade-offs in training data and the downstream effects on model performance. We then examine the effect of various components in the synthetic data pipeline on each data characteristic. This examination allows us to taxonomize and compare synthetic data generation algorithms through the components they utilize and the resulting effects on data QDC composition. This analysis extends into a discussion on the importance of balancing QDC in synthetic data for efficient reinforcement learning and self-improvement algorithms. Analogous to the QD trade-offs in training data, often there exist trade-offs between model output quality and output diversity which impact the composition of synthetic data. We observe that many models are currently evaluated and optimized only for output quality, thereby limiting output diversity and the potential for self-improvement. We argue that balancing these trade-offs is essential to the development of future self-improvement algorithms and highlight a number of works making progress in this direction.
♻ ☆ Order-Independence Without Fine Tuning NeurIPS 2024
The development of generative language models that can create long and coherent textual outputs via autoregression has lead to a proliferation of uses and a corresponding sweep of analyses as researches work to determine the limitations of this new paradigm. Unlike humans, these 'Large Language Models' (LLMs) are highly sensitive to small changes in their inputs, leading to unwanted inconsistency in their behavior. One problematic inconsistency when LLMs are used to answer multiple-choice questions or analyze multiple inputs is order dependency: the output of an LLM can (and often does) change significantly when sub-sequences are swapped, despite both orderings being semantically identical. In this paper we present Set-Based Prompting, a technique that guarantees the output of an LLM will not have order dependence on a specified set of sub-sequences. We show that this method provably eliminates order dependency, and that it can be applied to any transformer-based LLM to enable text generation that is unaffected by re-orderings. Delving into the implications of our method, we show that, despite our inputs being out of distribution, the impact on expected accuracy is small, where the expectation is over the order of uniformly chosen shuffling of the candidate responses, and usually significantly less in practice. Thus, Set-Based Prompting can be used as a 'dropped-in' method on fully trained models. Finally, we discuss how our method's success suggests that other strong guarantees can be obtained on LLM performance via modifying the input representations.
comment: 29 pages, 27 figures, Published in NeurIPS 2024 code https://github.com/reidmcy/set-based-prompting
♻ ☆ Systematic Review: Text Processing Algorithms in Machine Learning and Deep Learning for Mental Health Detection on Social Media
The global rise in depression necessitates innovative detection methods for early intervention. Social media provides a unique opportunity to identify depression through user-generated posts. This systematic review evaluates machine learning (ML) models for depression detection on social media, focusing on biases and methodological challenges throughout the ML lifecycle. A search of PubMed, IEEE Xplore, and Google Scholar identified 47 relevant studies published after 2010. The Prediction model Risk Of Bias ASsessment Tool (PROBAST) was utilized to assess methodological quality and risk of bias. Significant biases impacting model reliability and generalizability were found. There is a predominant reliance on Twitter (63.8%) and English-language content (over 90%), with most studies focusing on users from the United States and Europe. Non-probability sampling methods (approximately 80%) limit representativeness. Only 23% of studies explicitly addressed linguistic nuances like negations, crucial for accurate sentiment analysis. Inconsistent hyperparameter tuning was observed, with only 27.7% properly tuning models. About 17% did not adequately partition data into training, validation, and test sets, risking overfitting. While 74.5% used appropriate evaluation metrics for imbalanced data, others relied on accuracy without addressing class imbalance, potentially skewing results. Reporting transparency varied, often lacking critical methodological details. These findings highlight the need to diversify data sources, standardize preprocessing protocols, ensure consistent model development practices, address class imbalance, and enhance reporting transparency. By overcoming these challenges, future research can develop more robust and generalizable ML models for depression detection on social media, contributing to improved mental health outcomes globally.
♻ ☆ WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs NeurIPS 2024
We introduce WildGuard -- an open, light-weight moderation tool for LLM safety that achieves three goals: (1) identifying malicious intent in user prompts, (2) detecting safety risks of model responses, and (3) determining model refusal rate. Together, WildGuard serves the increasing needs for automatic safety moderation and evaluation of LLM interactions, providing a one-stop tool with enhanced accuracy and broad coverage across 13 risk categories. While existing open moderation tools such as Llama-Guard2 score reasonably well in classifying straightforward model interactions, they lag far behind a prompted GPT-4, especially in identifying adversarial jailbreaks and in evaluating models' refusals, a key measure for evaluating safety behaviors in model responses. To address these challenges, we construct WildGuardMix, a large-scale and carefully balanced multi-task safety moderation dataset with 92K labeled examples that cover vanilla (direct) prompts and adversarial jailbreaks, paired with various refusal and compliance responses. WildGuardMix is a combination of WildGuardTrain, the training data of WildGuard, and WildGuardTest, a high-quality human-annotated moderation test set with 5K labeled items covering broad risk scenarios. Through extensive evaluations on WildGuardTest and ten existing public benchmarks, we show that WildGuard establishes state-of-the-art performance in open-source safety moderation across all the three tasks compared to ten strong existing open-source moderation models (e.g., up to 26.4% improvement on refusal detection). Importantly, WildGuard matches and sometimes exceeds GPT-4 performance (e.g., up to 3.9% improvement on prompt harmfulness identification). WildGuard serves as a highly effective safety moderator in an LLM interface, reducing the success rate of jailbreak attacks from 79.8% to 2.4%.
comment: NeurIPS 2024 Camera Ready. First two authors contributed equally. Third and fourth authors contributed equally
♻ ☆ What Will My Model Forget? Forecasting Forgotten Examples in Language Model Refinement ICML 2024
Language models deployed in the wild make errors. However, simply updating the model with the corrected error instances causes catastrophic forgetting -- the updated model makes errors on instances learned during the instruction tuning or upstream training phase. Randomly replaying upstream data yields unsatisfactory performance and often comes with high variance and poor controllability. To this end, we try to forecast upstream examples that will be forgotten due to a model update for improved controllability of the replay process and interpretability. We train forecasting models given a collection of online learned examples and corresponding forgotten upstream pre-training examples. We propose a partially interpretable forecasting model based on the observation that changes in pre-softmax logit scores of pretraining examples resemble that of online learned examples, which performs decently on BART but fails on T5 models. We further show a black-box classifier based on inner products of example representations achieves better forecasting performance over a series of setups. Finally, we show that we reduce forgetting of upstream pretraining examples by replaying examples that are forecasted to be forgotten, demonstrating the practical utility of forecasting example forgetting.
comment: ICML 2024 (Spotlight)
Information Retrieval 14
☆ DEEPER: Dense Electroencephalography Passage Retrieval
Information retrieval systems have historically relied on explicit query formulation, requiring users to translate their information needs into text. This process is particularly disruptive during reading tasks, where users must interrupt their natural flow to formulate queries. We present DEEPER (Dense Electroencephalography Passage Retrieval), a novel framework that enables direct retrieval of relevant passages from users' neural signals during naturalistic reading without intermediate text translation. Building on dense retrieval architectures, DEEPER employs a dual-encoder approach with specialised components for processing neural data, mapping EEG signals and text passages into a shared semantic space. Through careful architecture design and cross-modal negative sampling strategies, our model learns to align neural patterns with their corresponding textual content. Experimental results on the ZuCo dataset demonstrate that direct brain-to-passage retrieval significantly outperforms current EEG-to-text baselines, achieving a 571% improvement in Precision@1. Our ablation studies reveal that the model successfully learns aligned representations between EEG and text modalities (0.29 cosine similarity), while our hard negative sampling strategy contributes to overall performance increases.
☆ Semantic Search and Recommendation Algorithm
This paper introduces a new semantic search algorithm that uses Word2Vec and Annoy Index to improve the efficiency of information retrieval from large datasets. The proposed approach addresses the limitations of traditional search methods by offering enhanced speed, accuracy, and scalability. Testing on datasets up to 100GB demonstrates the method's effectiveness in processing vast amounts of data while maintaining high precision and performance.
comment: 6 pages, 5 Figures
☆ PRECISE: Pre-training Sequential Recommenders with Collaborative and Semantic Information
Real-world recommendation systems commonly offer diverse content scenarios for users to interact with. Considering the enormous number of users in industrial platforms, it is infeasible to utilize a single unified recommendation model to meet the requirements of all scenarios. Usually, separate recommendation pipelines are established for each distinct scenario. This practice leads to challenges in comprehensively grasping users' interests. Recent research endeavors have been made to tackle this problem by pre-training models to encapsulate the overall interests of users. Traditional pre-trained recommendation models mainly capture user interests by leveraging collaborative signals. Nevertheless, a prevalent drawback of these systems is their incapacity to handle long-tail items and cold-start scenarios. With the recent advent of large language models, there has been a significant increase in research efforts focused on exploiting LLMs to extract semantic information for users and items. However, text-based recommendations highly rely on elaborate feature engineering and frequently fail to capture collaborative similarities. To overcome these limitations, we propose a novel pre-training framework for sequential recommendation, termed PRECISE. This framework combines collaborative signals with semantic information. Moreover, PRECISE employs a learning framework that initially models users' comprehensive interests across all recommendation scenarios and subsequently concentrates on the specific interests of target-scene behaviors. We demonstrate that PRECISE precisely captures the entire range of user interests and effectively transfers them to the target interests. Empirical findings reveal that the PRECISE framework attains outstanding performance on both public and industrial datasets.
☆ Methods for Legal Citation Prediction in the Age of LLMs: An Australian Law Case Study
In recent years, Large Language Models (LLMs) have shown great potential across a wide range of legal tasks. Despite these advances, mitigating hallucination remains a significant challenge, with state-of-the-art LLMs still frequently generating incorrect legal references. In this paper, we focus on the problem of legal citation prediction within the Australian law context, where correctly identifying and citing relevant legislations or precedents is critical. We compare several approaches: prompting general purpose and law-specialised LLMs, retrieval-only pipelines with both generic and domain-specific embeddings, task-specific instruction-tuning of LLMs, and hybrid strategies that combine LLMs with retrieval augmentation, query expansion, or voting ensembles. Our findings indicate that domain-specific pre-training alone is insufficient for achieving satisfactory citation accuracy even after law-specialised pre-training. In contrast, instruction tuning on our task-specific dataset dramatically boosts performance reaching the best results across all settings. We also highlight that database granularity along with the type of embeddings play a critical role in the performance of retrieval systems. Among retrieval-based approaches, hybrid methods consistently outperform retrieval-only setups, and among these, ensemble voting delivers the best result by combining the predictive quality of instruction-tuned LLMs with the retrieval system.
comment: For code, data, and models see https://auslawbench.github.io
☆ FM2DS: Few-Shot Multimodal Multihop Data Synthesis with Knowledge Distillation for Question Answering CVPR 2025
Multimodal multihop question answering is a complex task that requires reasoning over multiple sources of information, such as images and text, to answer questions. While there has been significant progress in visual question answering, the multihop setting remains unexplored due to the lack of high-quality datasets. Current methods focus on single-hop question answering or a single modality, which makes them unsuitable for real-world scenarios such as analyzing multimodal educational materials, summarizing lengthy academic articles, or interpreting scientific studies that combine charts, images, and text. To address this gap, we propose a novel methodology, introducing the first framework for creating a high-quality dataset that enables training models for multimodal multihop question answering. Our approach consists of a 5-stage pipeline that involves acquiring relevant multimodal documents from Wikipedia, synthetically generating high-level questions and answers, and validating them through rigorous criteria to ensure quality data. We evaluate our methodology by training models on our synthesized dataset and testing on two benchmarks, our results demonstrate that, with an equal sample size, models trained on our synthesized data outperform those trained on human-collected data by 1.9 in exact match (EM) on average. We believe our data synthesis method will serve as a strong foundation for training and evaluating multimodal multihop question answering models.
comment: 20 pages, 11 figures, 10 tables, Submitted to CVPR 2025
☆ CURE: Clinical Understanding & Retrieval Evaluation
Given the dominance of dense retrievers that do not generalize well beyond their training dataset distributions, domain-specific test sets are essential in evaluating retrieval. There are few test datasets for retrieval systems intended for use by healthcare providers in a point-of-care setting. To fill this gap we have collaborated with medical professionals to create CURE, an ad-hoc retrieval test dataset for passage ranking with 2000 queries spanning 10 medical domains with a monolingual (English) and two cross-lingual (French/Spanish -> English) conditions. In this paper, we describe how CURE was constructed and provide baseline results to showcase its effectiveness as an evaluation tool. CURE is published with a Creative Commons Attribution Non Commercial 4.0 license and can be accessed on Hugging Face.
☆ Bridging Conversational and Collaborative Signals for Conversational Recommendation
Conversational recommendation systems (CRS) leverage contextual information from conversations to generate recommendations but often struggle due to a lack of collaborative filtering (CF) signals, which capture user-item interaction patterns essential for accurate recommendations. We introduce Reddit-ML32M, a dataset that links reddit conversations with interactions on MovieLens 32M, to enrich item representations by leveraging collaborative knowledge and addressing interaction sparsity in conversational datasets. We propose an LLM-based framework that uses Reddit-ML32M to align LLM-generated recommendations with CF embeddings, refining rankings for better performance. We evaluate our framework against three sets of baselines: CF-based recommenders using only interactions from CRS tasks, traditional CRS models, and LLM-based methods relying on conversational context without item representations. Our approach achieves consistent improvements, including a 12.32% increase in Hit Rate and a 9.9% improvement in NDCG, outperforming the best-performing baseline that relies on conversational context but lacks collaborative item representations.
☆ Efficient user history modeling with amortized inference for deep learning recommendation models WWW 2025
We study user history modeling via Transformer encoders in deep learning recommendation models (DLRM). Such architectures can significantly improve recommendation quality, but usually incur high latency cost necessitating infrastructure upgrades or very small Transformer models. An important part of user history modeling is early fusion of the candidate item and various methods have been studied. We revisit early fusion and compare concatenation of the candidate to each history item against appending it to the end of the list as a separate item. Using the latter method, allows us to reformulate the recently proposed amortized history inference algorithm M-FALCON \cite{zhai2024actions} for the case of DLRM models. We show via experimental results that appending with cross-attention performs on par with concatenation and that amortization significantly reduces inference costs. We conclude with results from deploying this model on the LinkedIn Feed and Ads surfaces, where amortization reduces latency by 30\% compared to non-amortized inference.
comment: 5 pages, 3 figures, WWW 2025
♻ ☆ Predictive Models in Sequential Recommendations: Bridging Performance Laws with Data Quality Insights
Sequential Recommendation (SR) plays a critical role in predicting users' sequential preferences. Despite its growing prominence in various industries, the increasing scale of SR models incurs substantial computational costs and unpredictability, challenging developers to manage resources efficiently. Under this predicament, Scaling Laws have achieved significant success by examining the loss as models scale up. However, there remains a disparity between loss and model performance, which is of greater concern in practical applications. Moreover, as data continues to expand, it incorporates repetitive and inefficient data. In response, we introduce the Performance Law for SR models, which aims to theoretically investigate and model the relationship between model performance and data quality. Specifically, we first fit the HR and NDCG metrics to transformer-based SR models. Subsequently, we propose Approximate Entropy (ApEn) to assess data quality, presenting a more nuanced approach compared to traditional data quantity metrics. Our method enables accurate predictions across various dataset scales and model sizes, demonstrating a strong correlation in large SR models and offering insights into achieving optimal performance for any given model configuration.
comment: 12 pages, 5 figures
♻ ☆ Croissant: A Metadata Format for ML-Ready Datasets NeurIPS 2024
Data is a critical resource for machine learning (ML), yet working with data remains a key friction point. This paper introduces Croissant, a metadata format for datasets that creates a shared representation across ML tools, frameworks, and platforms. Croissant makes datasets more discoverable, portable, and interoperable, thereby addressing significant challenges in ML data management. Croissant is already supported by several popular dataset repositories, spanning hundreds of thousands of datasets, enabling easy loading into the most commonly-used ML frameworks, regardless of where the data is stored. Our initial evaluation by human raters shows that Croissant metadata is readable, understandable, complete, yet concise.
comment: Published at the NeurIPS 2024 Datasets and Benchmark Track. A shorter version appeared earlier in Proceedings of ACM SIGMOD/PODS'24 Data Management for End-to-End Machine Learning (DEEM) Workshop https://dl.acm.org/doi/10.1145/3650203.3663326
♻ ☆ Optimistic Query Routing in Clustering-based Approximate Maximum Inner Product Search
Clustering-based nearest neighbor search is an effective method in which points are partitioned into geometric shards to form an index, with only a few shards searched during query processing to find a set of top-$k$ vectors. Even though the search efficacy is heavily influenced by the algorithm that identifies the shards to probe, it has received little attention in the literature. This work bridges that gap by studying routing in clustering-based maximum inner product search. We unpack existing routers and notice the surprising contribution of optimism. We then take a page from the sequential decision making literature and formalize that insight following the principle of ``optimism in the face of uncertainty.'' In particular, we present a framework that incorporates the moments of the distribution of inner products within each shard to estimate the maximum inner product. We then present an instance of our algorithm that uses only the first two moments to reach the same accuracy as state-of-the-art routers such as ScaNN by probing up to $50\%$ fewer points on benchmark datasets. Our algorithm is also space-efficient: we design a sketch of the second moment whose size is independent of the number of points and requires $\mathcal{O}(1)$ vectors per shard.
♻ ☆ Enhancing Graph Contrastive Learning with Reliable and Informative Augmentation for Recommendation KDD 2025
Graph neural network(GNN) has been a powerful approach in collaborative filtering(CF) due to its ability to model high-order user-item relationships. Recently, to alleviate the data sparsity and enhance representation learning, many efforts have been conducted to integrate contrastive learning(CL) with GNNs. Despite the promising improvements, the contrastive view generation based on structure and representation perturbations in existing methods potentially disrupts the collaborative information in contrastive views, resulting in limited effectiveness of positive alignment. To overcome this issue, we propose CoGCL, a novel framework that aims to enhance graph contrastive learning by constructing contrastive views with stronger collaborative information via discrete codes. The core idea is to map users and items into discrete codes rich in collaborative information for reliable and informative contrastive view generation. To this end, we initially introduce a multi-level vector quantizer in an end-to-end manner to quantize user and item representations into discrete codes. Based on these discrete codes, we enhance the collaborative information of contrastive views by considering neighborhood structure and semantic relevance respectively. For neighborhood structure, we propose virtual neighbor augmentation by treating discrete codes as virtual neighbors, which expands an observed user-item interaction into multiple edges involving discrete codes. Regarding semantic relevance, we identify similar users/items based on shared discrete codes and interaction targets to generate the semantically relevant view. Through these strategies, we construct contrastive views with stronger collaborative information and develop a triple-view graph contrastive learning approach. Extensive experiments on four public datasets demonstrate the effectiveness of our proposed approach.
comment: Accepted by KDD 2025
♻ ☆ Enhancing FKG.in: automating Indian food composition analysis
This paper presents a novel approach to compute food composition data for Indian recipes using a knowledge graph for Indian food (FKG.in) and LLMs. The primary focus is to provide a broad overview of an automated food composition analysis workflow and describe its core functionalities: nutrition data aggregation, food composition analysis, and LLM-augmented information resolution. This workflow aims to complement FKG.in and iteratively supplement food composition data from verified knowledge bases. Additionally, this paper highlights the challenges of representing Indian food and accessing food composition data digitally. It also reviews three key sources of food composition data: the Indian Food Composition Tables, the Indian Nutrient Databank, and the Nutritionix API. Furthermore, it briefly outlines how users can interact with the workflow to obtain diet-based health recommendations and detailed food composition information for numerous recipes. We then explore the complex challenges of analyzing Indian recipe information across dimensions such as structure, multilingualism, and uncertainty as well as present our ongoing work on LLM-based solutions to address these issues. The methods proposed in this workshop paper for AI-driven knowledge curation and information resolution are application-agnostic, generalizable, and replicable for any domain.
comment: 15 pages, 5 figures, 30 references, International Conference on Pattern Recognition 2024 - Multimedia Assisted Dietary Management Workshop
♻ ☆ xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token
This paper introduces xRAG, an innovative context compression method tailored for retrieval-augmented generation. xRAG reinterprets document embeddings in dense retrieval--traditionally used solely for retrieval--as features from the retrieval modality. By employing a modality fusion methodology, xRAG seamlessly integrates these embeddings into the language model representation space, effectively eliminating the need for their textual counterparts and achieving an extreme compression rate. In xRAG, the only trainable component is the modality bridge, while both the retriever and the language model remain frozen. This design choice allows for the reuse of offline-constructed document embeddings and preserves the plug-and-play nature of retrieval augmentation. Experimental results demonstrate that xRAG achieves an average improvement of over 10% across six knowledge-intensive tasks, adaptable to various language model backbones, ranging from a dense 7B model to an 8x7B Mixture of Experts configuration. xRAG not only significantly outperforms previous context compression methods but also matches the performance of uncompressed models on several datasets, while reducing overall FLOPs by a factor of 3.53. Our work pioneers new directions in retrieval-augmented generation from the perspective of multimodality fusion, and we hope it lays the foundation for future efficient and scalable retrieval-augmented systems
comment: Neurips 2024
Machine Learning 150
☆ P3-PO: Prescriptive Point Priors for Visuo-Spatial Generalization of Robot Policies
Developing generalizable robot policies that can robustly handle varied environmental conditions and object instances remains a fundamental challenge in robot learning. While considerable efforts have focused on collecting large robot datasets and developing policy architectures to learn from such data, naively learning from visual inputs often results in brittle policies that fail to transfer beyond the training data. This work presents Prescriptive Point Priors for Policies or P3-PO, a novel framework that constructs a unique state representation of the environment leveraging recent advances in computer vision and robot learning to achieve improved out-of-distribution generalization for robot manipulation. This representation is obtained through two steps. First, a human annotator prescribes a set of semantically meaningful points on a single demonstration frame. These points are then propagated through the dataset using off-the-shelf vision models. The derived points serve as an input to state-of-the-art policy architectures for policy learning. Our experiments across four real-world tasks demonstrate an overall 43% absolute improvement over prior methods when evaluated in identical settings as training. Further, P3-PO exhibits 58% and 80% gains across tasks for new object instances and more cluttered environments respectively. Videos illustrating the robot's performance are best viewed at point-priors.github.io.
☆ Around the World in 80 Timesteps: A Generative Approach to Global Visual Geolocation
Global visual geolocation predicts where an image was captured on Earth. Since images vary in how precisely they can be localized, this task inherently involves a significant degree of ambiguity. However, existing approaches are deterministic and overlook this aspect. In this paper, we aim to close the gap between traditional geolocalization and modern generative methods. We propose the first generative geolocation approach based on diffusion and Riemannian flow matching, where the denoising process operates directly on the Earth's surface. Our model achieves state-of-the-art performance on three visual geolocation benchmarks: OpenStreetView-5M, YFCC-100M, and iNat21. In addition, we introduce the task of probabilistic visual geolocation, where the model predicts a probability distribution over all possible locations instead of a single point. We introduce new metrics and baselines for this task, demonstrating the advantages of our diffusion-based approach. Codes and models will be made available.
comment: Project page: https://nicolas-dufour.github.io/plonk
☆ Driv3R: Learning Dense 4D Reconstruction for Autonomous Driving
Realtime 4D reconstruction for dynamic scenes remains a crucial challenge for autonomous driving perception. Most existing methods rely on depth estimation through self-supervision or multi-modality sensor fusion. In this paper, we propose Driv3R, a DUSt3R-based framework that directly regresses per-frame point maps from multi-view image sequences. To achieve streaming dense reconstruction, we maintain a memory pool to reason both spatial relationships across sensors and dynamic temporal contexts to enhance multi-view 3D consistency and temporal integration. Furthermore, we employ a 4D flow predictor to identify moving objects within the scene to direct our network focus more on reconstructing these dynamic regions. Finally, we align all per-frame pointmaps consistently to the world coordinate system in an optimization-free manner. We conduct extensive experiments on the large-scale nuScenes dataset to evaluate the effectiveness of our method. Driv3R outperforms previous frameworks in 4D dynamic scene reconstruction, achieving 15x faster inference speed compared to methods requiring global alignment. Code: https://github.com/Barrybarry-Smith/Driv3R.
comment: Code is available at: https://github.com/Barrybarry-Smith/Driv3R
☆ Visual Lexicon: Rich Image Features in Language Space
We present Visual Lexicon, a novel visual language that encodes rich image information into the text space of vocabulary tokens while retaining intricate visual details that are often challenging to convey in natural language. Unlike traditional methods that prioritize either high-level semantics (e.g., CLIP) or pixel-level reconstruction (e.g., VAE), ViLex simultaneously captures rich semantic content and fine visual details, enabling high-quality image generation and comprehensive visual scene understanding. Through a self-supervised learning pipeline, ViLex generates tokens optimized for reconstructing input images using a frozen text-to-image (T2I) diffusion model, preserving the detailed information necessary for high-fidelity semantic-level reconstruction. As an image embedding in the language space, ViLex tokens leverage the compositionality of natural languages, allowing them to be used independently as "text tokens" or combined with natural language tokens to prompt pretrained T2I models with both visual and textual inputs, mirroring how we interact with vision-language models (VLMs). Experiments demonstrate that ViLex achieves higher fidelity in image reconstruction compared to text embeddings--even with a single ViLex token. Moreover, ViLex successfully performs various DreamBooth tasks in a zero-shot, unsupervised manner without fine-tuning T2I models. Additionally, ViLex serves as a powerful vision encoder, consistently improving vision-language model performance across 15 benchmarks relative to a strong SigLIP baseline.
comment: Tech report. 16 pages, 10 figures
☆ Proactive Agents for Multi-Turn Text-to-Image Generation Under Uncertainty
User prompts for generative AI models are often underspecified, leading to sub-optimal responses. This problem is particularly evident in text-to-image (T2I) generation, where users commonly struggle to articulate their precise intent. This disconnect between the user's vision and the model's interpretation often forces users to painstakingly and repeatedly refine their prompts. To address this, we propose a design for proactive T2I agents equipped with an interface to (1) actively ask clarification questions when uncertain, and (2) present their understanding of user intent as an understandable belief graph that a user can edit. We build simple prototypes for such agents and verify their effectiveness through both human studies and automated evaluation. We observed that at least 90% of human subjects found these agents and their belief graphs helpful for their T2I workflow. Moreover, we develop a scalable automated evaluation approach using two agents, one with a ground truth image and the other tries to ask as few questions as possible to align with the ground truth. On DesignBench, a benchmark we created for artists and designers, the COCO dataset (Lin et al., 2014), and ImageInWords (Garg et al., 2024), we observed that these T2I agents were able to ask informative questions and elicit crucial information to achieve successful alignment with at least 2 times higher VQAScore (Lin et al., 2024) than the standard single-turn T2I generation. Demo: https://github.com/google-deepmind/proactive_t2i_agents.
☆ Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
A key component of building safe and reliable language models is enabling the models to appropriately refuse to follow certain instructions or answer certain questions. We may want models to output refusal messages for various categories of user queries, for example, ill-posed questions, instructions for committing illegal acts, or queries which require information past the model's knowledge horizon. Engineering models that refuse to answer such questions is complicated by the fact that an individual may want their model to exhibit varying levels of sensitivity for refusing queries of various categories, and different users may want different refusal rates. The current default approach involves training multiple models with varying proportions of refusal messages from each category to achieve the desired refusal rates, which is computationally expensive and may require training a new model to accommodate each user's desired preference over refusal rates. To address these challenges, we propose refusal tokens, one such token for each refusal category or a single refusal token, which are prepended to the model's responses during training. We then show how to increase or decrease the probability of generating the refusal token for each category during inference to steer the model's refusal behavior. Refusal tokens enable controlling a single model's refusal rates without the need of any further fine-tuning, but only by selectively intervening during generation.
comment: 19 pages
☆ ONEBench to Test Them All: Sample-Level Benchmarking Over Open-Ended Capabilities
Traditional fixed test sets fall short in evaluating open-ended capabilities of foundation models. To address this, we propose ONEBench(OpeN-Ended Benchmarking), a new testing paradigm that consolidates individual evaluation datasets into a unified, ever-expanding sample pool. ONEBench allows users to generate custom, open-ended evaluation benchmarks from this pool, corresponding to specific capabilities of interest. By aggregating samples across test sets, ONEBench enables the assessment of diverse capabilities beyond those covered by the original test sets, while mitigating overfitting and dataset bias. Most importantly, it frames model evaluation as a collective process of selecting and aggregating sample-level tests. The shift from task-specific benchmarks to ONEBench introduces two challenges: (1)heterogeneity and (2)incompleteness. Heterogeneity refers to the aggregation over diverse metrics, while incompleteness describes comparing models evaluated on different data subsets. To address these challenges, we explore algorithms to aggregate sparse measurements into reliable model scores. Our aggregation algorithm ensures identifiability(asymptotically recovering ground-truth scores) and rapid convergence, enabling accurate model ranking with less data. On homogenous datasets, we show our aggregation algorithm provides rankings that highly correlate with those produced by average scores. We also demonstrate robustness to ~95% of measurements missing, reducing evaluation cost by up to 20x with little-to-no change in model rankings. We introduce ONEBench-LLM for language models and ONEBench-LMM for vision-language models, unifying evaluations across these domains. Overall, we present a technique for open-ended evaluation, which can aggregate over incomplete, heterogeneous sample-level measurements to continually grow a benchmark alongside the rapidly developing foundation models.
☆ Convolution goes higher-order: a biologically inspired mechanism empowers image classification
We propose a novel approach to image classification inspired by complex nonlinear biological visual processing, whereby classical convolutional neural networks (CNNs) are equipped with learnable higher-order convolutions. Our model incorporates a Volterra-like expansion of the convolution operator, capturing multiplicative interactions akin to those observed in early and advanced stages of biological visual processing. We evaluated this approach on synthetic datasets by measuring sensitivity to testing higher-order correlations and performance in standard benchmarks (MNIST, FashionMNIST, CIFAR10, CIFAR100 and Imagenette). Our architecture outperforms traditional CNN baselines, and achieves optimal performance with expansions up to 3rd/4th order, aligning remarkably well with the distribution of pixel intensities in natural images. Through systematic perturbation analysis, we validate this alignment by isolating the contributions of specific image statistics to model performance, demonstrating how different orders of convolution process distinct aspects of visual information. Furthermore, Representational Similarity Analysis reveals distinct geometries across network layers, indicating qualitatively different modes of visual information processing. Our work bridges neuroscience and deep learning, offering a path towards more effective, biologically inspired computer vision models. It provides insights into visual information processing and lays the groundwork for neural networks that better capture complex visual patterns, particularly in resource-constrained scenarios.
☆ Toward Non-Invasive Diagnosis of Bankart Lesions with Deep Learning SP
Bankart lesions, or anterior-inferior glenoid labral tears, are diagnostically challenging on standard MRIs due to their subtle imaging features-often necessitating invasive MRI arthrograms (MRAs). This study develops deep learning (DL) models to detect Bankart lesions on both standard MRIs and MRAs, aiming to improve diagnostic accuracy and reduce reliance on MRAs. We curated a dataset of 586 shoulder MRIs (335 standard, 251 MRAs) from 558 patients who underwent arthroscopy. Ground truth labels were derived from intraoperative findings, the gold standard for Bankart lesion diagnosis. Separate DL models for MRAs and standard MRIs were trained using the Swin Transformer architecture, pre-trained on a public knee MRI dataset. Predictions from sagittal, axial, and coronal views were ensembled to optimize performance. The models were evaluated on a 20% hold-out test set (117 MRIs: 46 MRAs, 71 standard MRIs). Bankart lesions were identified in 31.9% of MRAs and 8.6% of standard MRIs. The models achieved AUCs of 0.87 (86% accuracy, 83% sensitivity, 86% specificity) and 0.90 (85% accuracy, 82% sensitivity, 86% specificity) on standard MRIs and MRAs, respectively. These results match or surpass radiologist performance on our dataset and reported literature metrics. Notably, our model's performance on non-invasive standard MRIs matched or surpassed the radiologists interpreting MRAs. This study demonstrates the feasibility of using DL to address the diagnostic challenges posed by subtle pathologies like Bankart lesions. Our models demonstrate potential to improve diagnostic confidence, reduce reliance on invasive imaging, and enhance accessibility to care.
comment: Accepted for presentation at SPIE Medical Imaging 2025: Computer-Aided Diagnosis. The manuscript is expected to appear in the conference proceedings
☆ How to Merge Your Multimodal Models Over Time?
Model merging combines multiple expert models - finetuned from a base foundation model on diverse tasks and domains - into a single, more capable model. However, most existing model merging approaches assume that all experts are available simultaneously. In reality, new tasks and domains emerge progressively over time, requiring strategies to integrate the knowledge of expert models as they become available: a process we call temporal model merging. The temporal dimension introduces unique challenges not addressed in prior work, raising new questions such as: when training for a new task, should the expert model start from the merged past experts or from the original base model? Should we merge all models at each time step? Which merging techniques are best suited for temporal merging? Should different strategies be used to initialize the training and deploy the model? To answer these questions, we propose a unified framework called TIME - Temporal Integration of Model Expertise - which defines temporal model merging across three axes: (1) Initialization Phase, (2) Deployment Phase, and (3) Merging Technique. Using TIME, we study temporal model merging across model sizes, compute budgets, and learning horizons on the FoMo-in-Flux benchmark. Our comprehensive suite of experiments across TIME allows us to uncover key insights for temporal model merging, offering a better understanding of current challenges and best practices for effective temporal model merging.
comment: Technical Report. Code at https://github.com/ExplainableML/fomo_in_flux
☆ MISFEAT: Feature Selection for Subgroups with Systematic Missing Data
We investigate the problem of selecting features for datasets that can be naturally partitioned into subgroups (e.g., according to socio-demographic groups and age), each with its own dominant set of features. Within this subgroup-oriented framework, we address the challenge of systematic missing data, a scenario in which some feature values are missing for all tuples of a subgroup, due to flawed data integration, regulatory constraints, or privacy concerns. Feature selection is governed by finding mutual Information, a popular quantification of correlation, between features and a target variable. Our goal is to identify top-K feature subsets of some fixed size with the highest joint mutual information with a target variable. In the presence of systematic missing data, the closed form of mutual information could not simply be applied. We argue that in such a setting, leveraging relationships between available feature mutual information within a subgroup or across subgroups can assist inferring missing mutual information values. We propose a generalizable model based on heterogeneous graph neural network to identify interdependencies between feature-subgroup-target variable connections by modeling it as a multiplex graph, and employing information propagation between its nodes. We address two distinct scalability challenges related to training and propose principled solutions to tackle them. Through an extensive empirical evaluation, we demonstrate the efficacy of the proposed solutions both qualitatively and running time wise.
☆ OmniEvalKit: A Modular, Lightweight Toolbox for Evaluating Large Language Model and its Omni-Extensions
The rapid advancements in Large Language Models (LLMs) have significantly expanded their applications, ranging from multilingual support to domain-specific tasks and multimodal integration. In this paper, we present OmniEvalKit, a novel benchmarking toolbox designed to evaluate LLMs and their omni-extensions across multilingual, multidomain, and multimodal capabilities. Unlike existing benchmarks that often focus on a single aspect, OmniEvalKit provides a modular, lightweight, and automated evaluation system. It is structured with a modular architecture comprising a Static Builder and Dynamic Data Flow, promoting the seamless integration of new models and datasets. OmniEvalKit supports over 100 LLMs and 50 evaluation datasets, covering comprehensive evaluations across thousands of model-dataset combinations. OmniEvalKit is dedicated to creating an ultra-lightweight and fast-deployable evaluation framework, making downstream applications more convenient and versatile for the AI community.
☆ Impact of Privacy Parameters on Deep Learning Models for Image Classification
The project aims to develop differentially private deep learning models for image classification on CIFAR-10 datasets \cite{cifar10} and analyze the impact of various privacy parameters on model accuracy. We have implemented five different deep learning models, namely ConvNet, ResNet18, EfficientNet, ViT, and DenseNet121 and three supervised classifiers namely K-Nearest Neighbors, Naive Bayes Classifier and Support Vector Machine. We evaluated the performance of these models under varying settings. Our best performing model to date is EfficientNet with test accuracy of $59.63\%$ with the following parameters (Adam optimizer, batch size 256, epoch size 100, epsilon value 5.0, learning rate $1e-3$, clipping threshold 1.0, and noise multiplier 0.912).
comment: 10 pages
☆ Some Best Practices in Operator Learning
Hyperparameters searches are computationally expensive. This paper studies some general choices of hyperparameters and training methods specifically for operator learning. It considers the architectures DeepONets, Fourier neural operators and Koopman autoencoders for several differential equations to find robust trends. Some options considered are activation functions, dropout and stochastic weight averaging.
comment: arXiv admin note: substantial text overlap with arXiv:2412.04578
☆ Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone
Recent advances in learning decision-making policies can largely be attributed to training expressive policy models, largely via imitation learning. While imitation learning discards non-expert data, reinforcement learning (RL) can still learn from suboptimal data. However, instantiating RL training of a new policy class often presents a different challenge: most deep RL machinery is co-developed with assumptions on the policy class and backbone, resulting in poor performance when the policy class changes. For instance, SAC utilizes a low-variance reparameterization policy gradient for Gaussian policies, but this is unstable for diffusion policies and intractable for autoregressive categorical policies. To address this issue, we develop an offline RL and online fine-tuning approach called policy-agnostic RL (PA-RL) that can effectively train multiple policy classes, with varying architectures and sizes. We build off the basic idea that a universal supervised learning loss can replace the policy improvement step in RL, as long as it is applied on "optimized" actions. To obtain these optimized actions, we first sample multiple actions from a base policy, and run global optimization (i.e., re-ranking multiple action samples using the Q-function) and local optimization (i.e., running gradient steps on an action sample) to maximize the critic on these candidates. PA-RL enables fine-tuning diffusion and transformer policies with either autoregressive tokens or continuous action outputs, at different sizes, entirely via actor-critic RL. Moreover, PA-RL improves the performance and sample-efficiency by up to 2 times compared to existing offline RL and online fine-tuning methods. We show the first result that successfully fine-tunes OpenVLA, a 7B generalist robot policy, autonomously with Cal-QL, an online RL fine-tuning algorithm, improving from 40% to 70% in the real world in 40 minutes.
☆ Exploring Critical Testing Scenarios for Decision-Making Policies: An LLM Approach
Recent years have witnessed surprising achievements of decision-making policies across various fields, such as autonomous driving and robotics. Testing for decision-making policies is crucial with the existence of critical scenarios that may threaten their reliability. Numerous research efforts have been dedicated to testing these policies. However, there are still significant challenges, such as low testing efficiency and diversity due to the complexity of the policies and environments under test. Inspired by the remarkable capabilities of large language models (LLMs), in this paper, we propose an LLM-driven online testing framework for efficiently testing decision-making policies. The main idea is to employ an LLM-based test scenario generator to intelligently generate challenging test cases through contemplation and reasoning. Specifically, we first design a "generate-test-feedback" pipeline and apply templated prompt engineering to fully leverage the knowledge and reasoning abilities of LLMs. Then, we introduce a multi-scale scenario generation strategy to address the inherent challenges LLMs face in making fine adjustments, further enhancing testing efficiency. Finally, we evaluate the LLM-driven approach on five widely used benchmarks. The experimental results demonstrate that our method significantly outperforms baseline approaches in uncovering both critical and diverse scenarios.
comment: 16 pages, 13 figures
☆ I Don't Know: Explicit Modeling of Uncertainty with an [IDK] Token NeurIPS 2024
Large Language Models are known to capture real-world knowledge, allowing them to excel in many downstream tasks. Despite recent advances, these models are still prone to what are commonly known as hallucinations, causing them to emit unwanted and factually incorrect text. In this work, we propose a novel calibration method that can be used to combat hallucinations. We add a special [IDK] ("I don't know") token to the model's vocabulary and introduce an objective function that shifts probability mass to the [IDK] token for incorrect predictions. This approach allows the model to express uncertainty in its output explicitly. We evaluate our proposed method across multiple model architectures and factual downstream tasks. We find that models trained with our method are able to express uncertainty in places where they would previously make mistakes while suffering only a small loss of encoded knowledge. We further perform extensive ablation studies of multiple variations of our approach and provide a detailed analysis of the precision-recall tradeoff of our method.
comment: Published at NeurIPS 2024
☆ Off-Policy Maximum Entropy RL with Future State and Action Visitation Measures
We introduce a new maximum entropy reinforcement learning framework based on the distribution of states and actions visited by a policy. More precisely, an intrinsic reward function is added to the reward function of the Markov decision process that shall be controlled. For each state and action, this intrinsic reward is the relative entropy of the discounted distribution of states and actions (or features from these states and actions) visited during the next time steps. We first prove that an optimal exploration policy, which maximizes the expected discounted sum of intrinsic rewards, is also a policy that maximizes a lower bound on the state-action value function of the decision process under some assumptions. We also prove that the visitation distribution used in the intrinsic reward definition is the fixed point of a contraction operator. Following, we describe how to adapt existing algorithms to learn this fixed point and compute the intrinsic rewards to enhance exploration. A new practical off-policy maximum entropy reinforcement learning algorithm is finally introduced. Empirically, exploration policies have good state-action space coverage, and high-performing control policies are computed efficiently.
☆ Semantic Search and Recommendation Algorithm
This paper introduces a new semantic search algorithm that uses Word2Vec and Annoy Index to improve the efficiency of information retrieval from large datasets. The proposed approach addresses the limitations of traditional search methods by offering enhanced speed, accuracy, and scalability. Testing on datasets up to 100GB demonstrates the method's effectiveness in processing vast amounts of data while maintaining high precision and performance.
comment: 6 pages, 5 Figures
☆ The Narrow Gate: Localized Image-Text Communication in Vision-Language Models
Recent advances in multimodal training have significantly improved the integration of image understanding and generation within a unified model. This study investigates how vision-language models (VLMs) handle image-understanding tasks, specifically focusing on how visual information is processed and transferred to the textual domain. We compare VLMs that generate both images and text with those that output only text, highlighting key differences in information flow. We find that in models with multimodal outputs, image and text embeddings are more separated within the residual stream. Additionally, models vary in how information is exchanged from visual to textual tokens. VLMs that only output text exhibit a distributed communication pattern, where information is exchanged through multiple image tokens. In contrast, models trained for image and text generation rely on a single token that acts as a narrow gate for the visual information. We demonstrate that ablating this single token significantly deteriorates performance on image understanding tasks. Furthermore, modifying this token enables effective steering of the image semantics, showing that targeted, local interventions can reliably control the model's global behavior.
☆ Beyond Scalars: Concept-Based Alignment Analysis in Vision Transformers
Vision transformers (ViTs) can be trained using various learning paradigms, from fully supervised to self-supervised. Diverse training protocols often result in significantly different feature spaces, which are usually compared through alignment analysis. However, current alignment measures quantify this relationship in terms of a single scalar value, obscuring the distinctions between common and unique features in pairs of representations that share the same scalar alignment. We address this limitation by combining alignment analysis with concept discovery, which enables a breakdown of alignment into single concepts encoded in feature space. This fine-grained comparison reveals both universal and unique concepts across different representations, as well as the internal structure of concepts within each of them. Our methodological contributions address two key prerequisites for concept-based alignment: 1) For a description of the representation in terms of concepts that faithfully capture the geometry of the feature space, we define concepts as the most general structure they can possibly form - arbitrary manifolds, allowing hidden features to be described by their proximity to these manifolds. 2) To measure distances between concept proximity scores of two representations, we use a generalized Rand index and partition it for alignment between pairs of concepts. We confirm the superiority of our novel concept definition for alignment analysis over existing linear baselines in a sanity check. The concept-based alignment analysis of representations from four different ViTs reveals that increased supervision correlates with a reduction in the semantic structure of learned representations.
comment: 19 pages, 17 figures, code: https://github.com/jvielhaben/NLMCD-ALIGN
☆ PolytopeWalk: Sparse MCMC Sampling over Polytopes
High dimensional sampling is an important computational tool in statistics and other computational disciplines, with applications ranging from Bayesian statistical uncertainty quantification, metabolic modeling in systems biology to volume computation. We present $\textsf{PolytopeWalk}$, a new scalable Python library designed for uniform sampling over polytopes. The library provides an end-to-end solution, which includes preprocessing algorithms such as facial reduction and initialization methods. Six state-of-the-art MCMC algorithms on polytopes are implemented, including the Dikin, Vaidya, and John Walk. Additionally, we introduce novel sparse constrained formulations of these algorithms, enabling efficient sampling from sparse polytopes of the form $K_2 = \{x \in \mathbb{R}^d \ | \ Ax = b, x \succeq_k 0\}$. This implementation maintains sparsity in $A$, ensuring scalability to high dimensional settings $(d > 10^5)$. We demonstrate the improved sampling efficiency and per-iteration cost on both Netlib datasets and structured polytopes. $\textsf{PolytopeWalk}$ is available at github.com/ethz-randomwalk/polytopewalk with documentation at polytopewalk.readthedocs.io .
comment: 17 pages
☆ Copyright-Protected Language Generation via Adaptive Model Fusion
The risk of language models reproducing copyrighted material from their training data has led to the development of various protective measures. Among these, inference-time strategies that impose constraints via post-processing have shown promise in addressing the complexities of copyright regulation. However, they often incur prohibitive computational costs or suffer from performance trade-offs. To overcome these limitations, we introduce Copyright-Protecting Model Fusion (CP-Fuse), a novel approach that combines models trained on disjoint sets of copyrighted material during inference. In particular, CP-Fuse adaptively aggregates the model outputs to minimize the reproduction of copyrighted content, adhering to a crucial balancing property that prevents the regurgitation of memorized data. Through extensive experiments, we show that CP-Fuse significantly reduces the reproduction of protected material without compromising the quality of text and code generation. Moreover, its post-hoc nature allows seamless integration with other protective measures, further enhancing copyright safeguards. Lastly, we show that CP-Fuse is robust against common techniques for extracting training data.
comment: 47 pages, 21 Figures. arXiv admin note: substantial text overlap with arXiv:2407.20105
☆ AI TrackMate: Finally, Someone Who Will Give Your Music More Than Just "Sounds Great!" NeurIPS 2024
The rise of "bedroom producers" has democratized music creation, while challenging producers to objectively evaluate their work. To address this, we present AI TrackMate, an LLM-based music chatbot designed to provide constructive feedback on music productions. By combining LLMs' inherent musical knowledge with direct audio track analysis, AI TrackMate offers production-specific insights, distinguishing it from text-only approaches. Our framework integrates a Music Analysis Module, an LLM-Readable Music Report, and Music Production-Oriented Feedback Instruction, creating a plug-and-play, training-free system compatible with various LLMs and adaptable to future advancements. We demonstrate AI TrackMate's capabilities through an interactive web interface and present findings from a pilot study with a music producer. By bridging AI capabilities with the needs of independent producers, AI TrackMate offers on-demand analytical feedback, potentially supporting the creative process and skill development in music production. This system addresses the growing demand for objective self-assessment tools in the evolving landscape of independent music production.
comment: Accepted for the NeurIPS 2024 Creative AI Track
☆ Vulnerability of Text-Matching in ML/AI Conference Reviewer Assignments to Collusions
In the peer review process of top-tier machine learning (ML) and artificial intelligence (AI) conferences, reviewers are assigned to papers through automated methods. These assignment algorithms consider two main factors: (1) reviewers' expressed interests indicated by their bids for papers, and (2) reviewers' domain expertise inferred from the similarity between the text of their previously published papers and the submitted manuscripts. A significant challenge these conferences face is the existence of collusion rings, where groups of researchers manipulate the assignment process to review each other's papers, providing positive evaluations regardless of their actual quality. Most efforts to combat collusion rings have focused on preventing bid manipulation, under the assumption that the text similarity component is secure. In this paper, we demonstrate that even in the absence of bidding, colluding reviewers and authors can exploit the machine learning based text-matching component of reviewer assignment used at top ML/AI venues to get assigned their target paper. We also highlight specific vulnerabilities within this system and offer suggestions to enhance its robustness.
☆ VOPy: A Framework for Black-box Vector Optimization
We introduce VOPy, an open-source Python library designed to address black-box vector optimization, where multiple objectives must be optimized simultaneously with respect to a partial order induced by a convex cone. VOPy extends beyond traditional multi-objective optimization (MOO) tools by enabling flexible, cone-based ordering of solutions; with an application scope that includes environments with observation noise, discrete or continuous design spaces, limited budgets, and batch observations. VOPy provides a modular architecture, facilitating the integration of existing methods and the development of novel algorithms. We detail VOPy's architecture, usage, and potential to advance research and application in the field of vector optimization. The source code for VOPy is available at https://github.com/Bilkent-CYBORG/VOPy.
☆ Towards Controllable Speech Synthesis in the Era of Large Language Models: A Survey
Text-to-speech (TTS), also known as speech synthesis, is a prominent research area that aims to generate natural-sounding human speech from text. Recently, with the increasing industrial demand, TTS technologies have evolved beyond synthesizing human-like speech to enabling controllable speech generation. This includes fine-grained control over various attributes of synthesized speech such as emotion, prosody, timbre, and duration. Besides, advancements in deep learning, such as diffusion and large language models, have significantly enhanced controllable TTS over the past several years. In this paper, we conduct a comprehensive survey of controllable TTS, covering approaches ranging from basic control techniques to methods utilizing natural language prompts, aiming to provide a clear understanding of the current state of research. We examine the general controllable TTS pipeline, challenges, model architectures, and control strategies, offering a comprehensive and clear taxonomy of existing methods. Additionally, we provide a detailed summary of datasets and evaluation metrics and shed some light on the applications and future directions of controllable TTS. To the best of our knowledge, this survey paper provides the first comprehensive review of emerging controllable TTS methods, which can serve as a beneficial resource for both academic researchers and industry practitioners.
comment: A comprehensive survey on controllable TTS, 23 pages, 6 tables, 4 figures, 280 references
☆ Self-Interested Agents in Collaborative Learning: An Incentivized Adaptive Data-Centric Framework
We propose a framework for adaptive data-centric collaborative learning among self-interested agents, coordinated by an arbiter. Designed to handle the incremental nature of real-world data, the framework operates in an online manner: at each step, the arbiter collects a batch of data from agents, trains a machine learning model, and provides each agent with a distinct model reflecting its data contributions. This setup establishes a feedback loop where shared data influence model updates, and the resulting models guide future data-sharing strategies. Agents evaluate and partition their data, selecting a partition to share using a stochastic parameterized policy optimized via policy gradient methods to optimize the utility of the received model as defined by agent-specific evaluation functions. On the arbiter side, the expected loss function over the true data distribution is optimized, incorporating agent-specific weights to account for distributional differences arising from diverse sources and selective sharing. A bilevel optimization algorithm jointly learns the model parameters and agent-specific weights. Mean-zero noise, computed using a distortion function that adjusts these agent-specific weights, is introduced to generate distinct agent-specific models, promoting valuable data sharing without requiring separate training. Our framework is underpinned by non-asymptotic analyses, ensuring convergence of the agent-side policy optimization to an approximate stationary point of the evaluation functions and convergence of the arbiter-side optimization to an approximate stationary point of the expected loss function.
☆ CONDEN-FI: Consistency and Diversity Learning-based Multi-View Unsupervised Feature and In-stance Co-Selection
The objective of multi-view unsupervised feature and instance co-selection is to simultaneously iden-tify the most representative features and samples from multi-view unlabeled data, which aids in mit-igating the curse of dimensionality and reducing instance size to improve the performance of down-stream tasks. However, existing methods treat feature selection and instance selection as two separate processes, failing to leverage the potential interactions between the feature and instance spaces. Addi-tionally, previous co-selection methods for multi-view data require concatenating different views, which overlooks the consistent information among them. In this paper, we propose a CONsistency and DivErsity learNing-based multi-view unsupervised Feature and Instance co-selection (CONDEN-FI) to address the above-mentioned issues. Specifically, CONDEN-FI reconstructs mul-ti-view data from both the sample and feature spaces to learn representations that are consistent across views and specific to each view, enabling the simultaneous selection of the most important features and instances. Moreover, CONDEN-FI adaptively learns a view-consensus similarity graph to help select both dissimilar and similar samples in the reconstructed data space, leading to a more diverse selection of instances. An efficient algorithm is developed to solve the resultant optimization problem, and the comprehensive experimental results on real-world datasets demonstrate that CONDEN-FI is effective compared to state-of-the-art methods.
☆ DEX: Data Channel Extension for Efficient CNN Inference on Tiny AI Accelerators NeurIPS 2024
Tiny machine learning (TinyML) aims to run ML models on small devices and is increasingly favored for its enhanced privacy, reduced latency, and low cost. Recently, the advent of tiny AI accelerators has revolutionized the TinyML field by significantly enhancing hardware processing power. These accelerators, equipped with multiple parallel processors and dedicated per-processor memory instances, offer substantial performance improvements over traditional microcontroller units (MCUs). However, their limited data memory often necessitates downsampling input images, resulting in accuracy degradation. To address this challenge, we propose Data channel EXtension (DEX), a novel approach for efficient CNN execution on tiny AI accelerators. DEX incorporates additional spatial information from original images into input images through patch-wise even sampling and channel-wise stacking, effectively extending data across input channels. By leveraging underutilized processors and data memory for channel extension, DEX facilitates parallel execution without increasing inference latency. Our evaluation with four models and four datasets on tiny AI accelerators demonstrates that this simple idea improves accuracy on average by 3.5%p while keeping the inference latency the same on the AI accelerator. The source code is available at https://github.com/Nokia-Bell-Labs/data-channel-extension.
comment: Accepted to NeurIPS 2024
☆ ProcessBench: Identifying Process Errors in Mathematical Reasoning
As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.
☆ When Dimensionality Reduction Meets Graph (Drawing) Theory: Introducing a Common Framework, Challenges and Opportunities
In the vast landscape of visualization research, Dimensionality Reduction (DR) and graph analysis are two popular subfields, often essential to most visual data analytics setups. DR aims to create representations to support neighborhood and similarity analysis on complex, large datasets. Graph analysis focuses on identifying the salient topological properties and key actors within networked data, with specialized research on investigating how such features could be presented to the user to ease the comprehension of the underlying structure. Although these two disciplines are typically regarded as disjoint subfields, we argue that both fields share strong similarities and synergies that can potentially benefit both. Therefore, this paper discusses and introduces a unifying framework to help bridge the gap between DR and graph (drawing) theory. Our goal is to use the strongly math-grounded graph theory to improve the overall process of creating DR visual representations. We propose how to break the DR process into well-defined stages, discussing how to match some of the DR state-of-the-art techniques to this framework and presenting ideas on how graph drawing, topology features, and some popular algorithms and strategies used in graph analysis can be employed to improve DR topology extraction, embedding generation, and result validation. We also discuss the challenges and identify opportunities for implementing and using our framework, opening directions for future visualization research.
☆ Prediction of Occluded Pedestrians in Road Scenes using Human-like Reasoning: Insights from the OccluRoads Dataset
Pedestrian detection is a critical task in autonomous driving, aimed at enhancing safety and reducing risks on the road. Over recent years, significant advancements have been made in improving detection performance. However, these achievements still fall short of human perception, particularly in cases involving occluded pedestrians, especially entirely invisible ones. In this work, we present the Occlusion-Rich Road Scenes with Pedestrians (OccluRoads) dataset, which features a diverse collection of road scenes with partially and fully occluded pedestrians in both real and virtual environments. All scenes are meticulously labeled and enriched with contextual information that encapsulates human perception in such scenarios. Using this dataset, we developed a pipeline to predict the presence of occluded pedestrians, leveraging Knowledge Graph (KG), Knowledge Graph Embedding (KGE), and a Bayesian inference process. Our approach achieves a F1 score of 0.91, representing an improvement of up to 42% compared to traditional machine learning models.
☆ On How Iterative Magnitude Pruning Discovers Local Receptive Fields in Fully Connected Neural Networks
Since its use in the Lottery Ticket Hypothesis, iterative magnitude pruning (IMP) has become a popular method for extracting sparse subnetworks that can be trained to high performance. Despite this, the underlying nature of IMP's general success remains unclear. One possibility is that IMP is especially capable of extracting and maintaining strong inductive biases. In support of this, recent work has shown that applying IMP to fully connected neural networks (FCNs) leads to the emergence of local receptive fields (RFs), an architectural feature present in mammalian visual cortex and convolutional neural networks. The question of how IMP is able to do this remains unanswered. Inspired by results showing that training FCNs on synthetic images with highly non-Gaussian statistics (e.g., sharp edges) is sufficient to drive the formation of local RFs, we hypothesize that IMP iteratively maximizes the non-Gaussian statistics present in the representations of FCNs, creating a feedback loop that enhances localization. We develop a new method for measuring the effect of individual weights on the statistics of the FCN representations ("cavity method"), which allows us to find evidence in support of this hypothesis. Our work, which is the first to study the effect IMP has on the representations of neural networks, sheds parsimonious light one way in which IMP can drive the formation of strong inductive biases.
comment: 10 pages, 5 figures, comments welcome!
☆ Sloth: scaling laws for LLM skills to predict multi-benchmark performance across families
Scaling laws for large language models (LLMs) predict model performance based on parameters like size and training data. However, differences in training configurations and data processing across model families lead to significant variations in benchmark performance, making it difficult for a single scaling law to generalize across all LLMs. On the other hand, training family-specific scaling laws requires training models of varying sizes for every family. In this work, we propose Skills Scaling Laws (SSLaws, pronounced as Sloth), a novel scaling law that leverages publicly available benchmark data and assumes LLM performance is driven by low-dimensional latent skills, such as reasoning and instruction following. These latent skills are influenced by computational resources like model size and training tokens but with varying efficiencies across model families. Sloth exploits correlations across benchmarks to provide more accurate and interpretable predictions while alleviating the need to train multiple LLMs per family. We present both theoretical results on parameter identification and empirical evaluations on 12 prominent benchmarks, from Open LLM Leaderboard v1/v2, demonstrating that Sloth predicts LLM performance efficiently and offers insights into scaling behaviors for downstream tasks such as coding and emotional intelligence applications.
☆ Understanding Factual Recall in Transformers via Associative Memories
Large language models have demonstrated an impressive ability to perform factual recall. Prior work has found that transformers trained on factual recall tasks can store information at a rate proportional to their parameter count. In our work, we show that shallow transformers can use a combination of associative memories to obtain such near optimal storage capacity. We begin by proving that the storage capacities of both linear and MLP associative memories scale linearly with parameter count. We next introduce a synthetic factual recall task, and prove that a transformer with a single layer of self-attention followed by an MLP can obtain 100% accuracy on the task whenever either the total number of self-attention parameters or MLP parameters scales (up to log factors) linearly with the number of facts. In particular, the transformer can trade off between using the value matrices or the MLP as an associative memory to store the dataset of facts. We complement these expressivity results with an analysis of the gradient flow trajectory of a simplified linear attention model trained on our factual recall task, where we show that the model exhibits sequential learning behavior.
☆ Inverting Visual Representations with Detection Transformers
Understanding the mechanisms underlying deep neural networks in computer vision remains a fundamental challenge. While many prior approaches have focused on visualizing intermediate representations within deep neural networks, particularly convolutional neural networks, these techniques have yet to be thoroughly explored in transformer-based vision models. In this study, we apply the approach of training inverse models to reconstruct input images from intermediate layers within a Detection Transformer, showing that this approach is efficient and feasible for transformer-based vision models. Through qualitative and quantitative evaluations of reconstructed images across model stages, we demonstrate critical properties of Detection Transformers, including contextual shape preservation, inter-layer correlation, and robustness to color perturbations, illustrating how these characteristics emerge within the model's architecture. Our findings contribute to a deeper understanding of transformer-based vision models. The code for reproducing our experiments will be made available at github.com/wiskott-lab/inverse-detection-transformer.
☆ Unraveling the Complexity of Memory in RL Agents: an Approach for Classification and Evaluation
The incorporation of memory into agents is essential for numerous tasks within the domain of Reinforcement Learning (RL). In particular, memory is paramount for tasks that require the utilization of past information, adaptation to novel environments, and improved sample efficiency. However, the term ``memory'' encompasses a wide range of concepts, which, coupled with the lack of a unified methodology for validating an agent's memory, leads to erroneous judgments about agents' memory capabilities and prevents objective comparison with other memory-enhanced agents. This paper aims to streamline the concept of memory in RL by providing practical precise definitions of agent memory types, such as long-term versus short-term memory and declarative versus procedural memory, inspired by cognitive science. Using these definitions, we categorize different classes of agent memory, propose a robust experimental methodology for evaluating the memory capabilities of RL agents, and standardize evaluations. Furthermore, we empirically demonstrate the importance of adhering to the proposed methodology when evaluating different types of agent memory by conducting experiments with different RL agents and what its violation leads to.
comment: 18 pages, 6 figures
☆ BATseg: Boundary-aware Multiclass Spinal Cord Tumor Segmentation on 3D MRI Scans ECCV 2024
Spinal cord tumors significantly contribute to neurological morbidity and mortality. Precise morphometric quantification, encompassing the size, location, and type of such tumors, holds promise for optimizing treatment planning strategies. Although recent methods have demonstrated excellent performance in medical image segmentation, they primarily focus on discerning shapes with relatively large morphology such as brain tumors, ignoring the challenging problem of identifying spinal cord tumors which tend to have tiny sizes, diverse locations, and shapes. To tackle this hard problem of multiclass spinal cord tumor segmentation, we propose a new method, called BATseg, to learn a tumor surface distance field by applying our new multiclass boundary-aware loss function. To verify the effectiveness of our approach, we also introduce the first and large-scale spinal cord tumor dataset. It comprises gadolinium-enhanced T1-weighted 3D MRI scans from 653 patients and contains the four most common spinal cord tumor types: astrocytomas, ependymomas, hemangioblastomas, and spinal meningiomas. Extensive experiments on our dataset and another public kidney tumor segmentation dataset show that our proposed method achieves superior performance for multiclass tumor segmentation.
comment: ECCV 2024 Workshop on BioImage Computing. Code and data are available at: https://github.com/vLAR-group/BATseg
☆ A cautionary tale on the cost-effectiveness of collaborative AI in real-world medical applications
Background. Federated learning (FL) has gained wide popularity as a collaborative learning paradigm enabling collaborative AI in sensitive healthcare applications. Nevertheless, the practical implementation of FL presents technical and organizational challenges, as it generally requires complex communication infrastructures. In this context, consensus-based learning (CBL) may represent a promising collaborative learning alternative, thanks to the ability of combining local knowledge into a federated decision system, while potentially reducing deployment overhead. Methods. In this work we propose an extensive benchmark of the accuracy and cost-effectiveness of a panel of FL and CBL methods in a wide range of collaborative medical data analysis scenarios. The benchmark includes 7 different medical datasets, encompassing 3 machine learning tasks, 8 different data modalities, and multi-centric settings involving 3 to 23 clients. Findings. Our results reveal that CBL is a cost-effective alternative to FL. When compared across the panel of medical dataset in the considered benchmark, CBL methods provide equivalent accuracy to the one achieved by FL.Nonetheless, CBL significantly reduces training time and communication cost (resp. 15 fold and 60 fold decrease) (p < 0.05). Interpretation. This study opens a novel perspective on the deployment of collaborative AI in real-world applications, whereas the adoption of cost-effective methods is instrumental to achieve sustainability and democratisation of AI by alleviating the need for extensive computational resources.
☆ Improving text-conditioned latent diffusion for cancer pathology
The development of generative models in the past decade has allowed for hyperrealistic data synthesis. While potentially beneficial, this synthetic data generation process has been relatively underexplored in cancer histopathology. One algorithm for synthesising a realistic image is diffusion; it iteratively converts an image to noise and learns the recovery process from this noise [Wang and Vastola, 2023]. While effective, it is highly computationally expensive for high-resolution images, rendering it infeasible for histopathology. The development of Variational Autoencoders (VAEs) has allowed us to learn the representation of complex high-resolution images in a latent space. A vital by-product of this is the ability to compress high-resolution images to space and recover them lossless. The marriage of diffusion and VAEs allows us to carry out diffusion in the latent space of an autoencoder, enabling us to leverage the realistic generative capabilities of diffusion while maintaining reasonable computational requirements. Rombach et al. [2021b] and Yellapragada et al. [2023] build foundational models for this task, paving the way to generate realistic histopathology images. In this paper, we discuss the pitfalls of current methods, namely [Yellapragada et al., 2023] and resolve critical errors while proposing improvements along the way. Our methods achieve an FID score of 21.11, beating its SOTA counterparts in [Yellapragada et al., 2023] by 1.2 FID, while presenting a train-time GPU memory usage reduction of 7%.
☆ SimuDICE: Offline Policy Optimization Through World Model Updates and DICE Estimation
In offline reinforcement learning, deriving an effective policy from a pre-collected set of experiences is challenging due to the distribution mismatch between the target policy and the behavioral policy used to collect the data, as well as the limited sample size. Model-based reinforcement learning improves sample efficiency by generating simulated experiences using a learned dynamic model of the environment. However, these synthetic experiences often suffer from the same distribution mismatch. To address these challenges, we introduce SimuDICE, a framework that iteratively refines the initial policy derived from offline data using synthetically generated experiences from the world model. SimuDICE enhances the quality of these simulated experiences by adjusting the sampling probabilities of state-action pairs based on stationary DIstribution Correction Estimation (DICE) and the estimated confidence in the model's predictions. This approach guides policy improvement by balancing experiences similar to those frequently encountered with ones that have a distribution mismatch. Our experiments show that SimuDICE achieves performance comparable to existing algorithms while requiring fewer pre-collected experiences and planning steps, and it remains robust across varying data collection policies.
comment: Published at BNAIC/BeNeLearn 2024
☆ An inferential measure of dependence between two systems using Bayesian model comparison
We propose to quantify dependence between two systems $X$ and $Y$ in a dataset $D$ based on the Bayesian comparison of two models: one, $H_0$, of statistical independence and another one, $H_1$, of dependence. In this framework, dependence between $X$ and $Y$ in $D$, denoted $B(X,Y|D)$, is quantified as $P(H_1|D)$, the posterior probability for the model of dependence given $D$, or any strictly increasing function thereof. It is therefore a measure of the evidence for dependence between $X$ and $Y$ as modeled by $H_1$ and observed in $D$. We review several statistical models and reconsider standard results in the light of $B(X,Y|D)$ as a measure of dependence. Using simulations, we focus on two specific issues: the effect of noise and the behavior of $B(X,Y|D)$ when $H_1$ has a parameter coding for the intensity of dependence. We then derive some general properties of $B(X,Y|D)$, showing that it quantifies the information contained in $D$ in favor of $H_1$ versus $H_0$. While some of these properties are typical of what is expected from a valid measure of dependence, others are novel and naturally appear as desired features for specific measures of dependence, which we call inferential. We finally put these results in perspective; in particular, we discuss the consequences of using the Bayesian framework as well as the similarities and differences between $B(X,Y|D)$ and mutual information.
comment: To be published in IEEE Transaction on Systems, Man, and Cybernetics: Systems
☆ From Uncertainty to Trust: Enhancing Reliability in Vision-Language Models with Uncertainty-Guided Dropout Decoding
Large vision-language models (LVLMs) demonstrate remarkable capabilities in multimodal tasks but are prone to misinterpreting visual inputs, often resulting in hallucinations and unreliable outputs. To address these challenges, we propose Dropout Decoding, a novel inference-time approach that quantifies the uncertainty of visual tokens and selectively masks uncertain tokens to improve decoding. Our method measures the uncertainty of each visual token by projecting it onto the text space and decomposing it into aleatoric and epistemic components. Specifically, we focus on epistemic uncertainty, which captures perception-related errors more effectively. Inspired by dropout regularization, we introduce uncertainty-guided token dropout, which applies the dropout principle to input visual tokens instead of model parameters, and during inference rather than training. By aggregating predictions from an ensemble of masked decoding contexts, Dropout Decoding robustly mitigates errors arising from visual token misinterpretations. Evaluations on benchmarks including CHAIR, THRONE, and MMBench demonstrate that Dropout Decoding significantly reduces object hallucinations (OH) and enhances both reliability and quality of LVLM outputs across diverse visual contexts.
comment: Code is released at https://github.com/kigb/DropoutDecoding
☆ Food for thought: How can machine learning help better predict and understand changes in food prices?
In this work, we address a lack of systematic understanding of fluctuations in food affordability in Canada. Canada's Food Price Report (CPFR) is an annual publication that predicts food inflation over the next calendar year. The published predictions are a collaborative effort between forecasting teams that each employ their own approach at Canadian Universities: Dalhousie University, the University of British Columbia, the University of Saskatchewan, and the University of Guelph/Vector Institute. While the University of Guelph/Vector Institute forecasting team has leveraged machine learning (ML) in previous reports, the most recent editions (2024--2025) have also included a human-in-the-loop approach. For the 2025 report, this focus was expanded to evaluate several different data-centric approaches to improve forecast accuracy. In this study, we evaluate how different types of forecasting models perform when estimating food price fluctuations. We also examine the sensitivity of models that curate time series data representing key factors in food pricing.
☆ Active Learning with Context Sampling and One-vs-Rest Entropy for Semantic Segmentation WACV 2025
Multi-class semantic segmentation remains a cornerstone challenge in computer vision. Yet, dataset creation remains excessively demanding in time and effort, especially for specialized domains. Active Learning (AL) mitigates this challenge by selecting data points for annotation strategically. However, existing patch-based AL methods often overlook boundary pixels critical information, essential for accurate segmentation. We present OREAL, a novel patch-based AL method designed for multi-class semantic segmentation. OREAL enhances boundary detection by employing maximum aggregation of pixel-wise uncertainty scores. Additionally, we introduce one-vs-rest entropy, a novel uncertainty score function that computes class-wise uncertainties while achieving implicit class balancing during dataset creation. Comprehensive experiments across diverse datasets and model architectures validate our hypothesis.
comment: WACV 2025, 8 pages
☆ Gated Delta Networks: Improving Mamba2 with Delta Rule
Linear Transformers have gained attention as efficient alternatives to standard Transformers, but their performance in retrieval and long-context tasks has been limited. To address these limitations, recent work has explored two distinct mechanisms: gating for adaptive memory control and the delta update rule for precise memory modifications. We observe that these mechanisms are complementary: gating enables rapid memory erasure while the delta rule facilitates targeted updates. Building on this insight, we introduce the gated delta rule and develop a parallel training algorithm optimized for modern hardware. Our proposed architecture, Gated DeltaNet, consistently surpasses existing models like Mamba2 and DeltaNet across multiple benchmarks, including language modeling, common-sense reasoning, in-context retrieval, length extrapolation, and long-context understanding. We further enhance performance by developing hybrid architectures that combine Gated DeltaNet layers with sliding window attention or Mamba2 layers, achieving both improved training efficiency and superior task performance.
comment: Preprint
☆ How Certain are Uncertainty Estimates? Three Novel Earth Observation Datasets for Benchmarking Uncertainty Quantification in Machine Learning
Uncertainty quantification (UQ) is essential for assessing the reliability of Earth observation (EO) products. However, the extensive use of machine learning models in EO introduces an additional layer of complexity, as those models themselves are inherently uncertain. While various UQ methods do exist for machine learning models, their performance on EO datasets remains largely unevaluated. A key challenge in the community is the absence of the ground truth for uncertainty, i.e. how certain the uncertainty estimates are, apart from the labels for the image/signal. This article fills this gap by introducing three benchmark datasets specifically designed for UQ in EO machine learning models. These datasets address three common problem types in EO: regression, image segmentation, and scene classification. They enable a transparent comparison of different UQ methods for EO machine learning models. We describe the creation and characteristics of each dataset, including data sources, preprocessing steps, and label generation, with a particular focus on calculating the reference uncertainty. We also showcase baseline performance of several machine learning models on each dataset, highlighting the utility of these benchmarks for model development and comparison. Overall, this article offers a valuable resource for researchers and practitioners working in artificial intelligence for EO, promoting a more accurate and reliable quality measure of the outputs of machine learning models. The dataset and code are accessible via https://gitlab.lrz.de/ai4eo/WG_Uncertainty.
comment: Submitted to IEEE Geoscience and Remote Sensing Magazine
☆ Echocardiography to Cardiac MRI View Transformation for Real-Time Blind Restoration
Echocardiography is the most widely used imaging to monitor cardiac functions, serving as the first line in early detection of myocardial ischemia and infarction. However, echocardiography often suffers from several artifacts including sensor noise, lack of contrast, severe saturation, and missing myocardial segments which severely limit its usage in clinical diagnosis. In recent years, several machine learning methods have been proposed to improve echocardiography views. Yet, these methods usually address only a specific problem (e.g. denoising) and thus cannot provide a robust and reliable restoration in general. On the other hand, cardiac MRI provides a clean view of the heart without suffering such severe issues. However, due to its significantly higher cost, it is often only afforded by a few major hospitals, hence hindering its use and accessibility. In this pilot study, we propose a novel approach to transform echocardiography into the cardiac MRI view. For this purpose, Echo2MRI dataset, consisting of echocardiography and real cardiac MRI image pairs, is composed and will be shared publicly. A dedicated Cycle-consistent Generative Adversarial Network (Cycle-GAN) is trained to learn the transformation from echocardiography frames to cardiac MRI views. An extensive set of qualitative evaluations shows that the proposed transformer can synthesize high-quality artifact-free synthetic cardiac MRI views from a given sequence of echocardiography frames. Medical evaluations performed by a group of cardiologists further demonstrate that synthetic MRI views are indistinguishable from their original counterparts and are preferred over their initial sequence of echocardiography frames for diagnosis in 78.9% of the cases.
comment: 18 pages, 42 figures
☆ Can foundation models actively gather information in interactive environments to test hypotheses?
While problem solving is a standard evaluation task for foundation models, a crucial component of problem solving -- actively and strategically gathering information to test hypotheses -- has not been closely investigated. To assess the information gathering abilities of foundation models in interactive environments, we introduce a framework in which a model must determine the factors influencing a hidden reward function by iteratively reasoning about its previously gathered information and proposing its next exploratory action to maximize information gain at each step. We implement this framework in both a text-based environment, which offers a tightly controlled setting and enables high-throughput parameter sweeps, and in an embodied 3D environment, which requires addressing complexities of multi-modal interaction more relevant to real-world applications. We further investigate whether approaches such as self-correction and increased inference time improve information gathering efficiency. In a relatively simple task that requires identifying a single rewarding feature, we find that LLM's information gathering capability is close to optimal. However, when the model must identify a conjunction of rewarding features, performance is suboptimal. The hit in performance is due partly to the model translating task description to a policy and partly to the model's effectiveness in using its in-context memory. Performance is comparable in both text and 3D embodied environments, although imperfect visual object recognition reduces its accuracy in drawing conclusions from gathered information in the 3D embodied case. For single-feature-based rewards, we find that smaller models curiously perform better; for conjunction-based rewards, incorporating self correction into the model improves performance.
☆ An Adaptively Inexact Method for Bilevel Learning Using Primal-Dual Style Differentiation
We consider a bilevel learning framework for learning linear operators. In this framework, the learnable parameters are optimized via a loss function that also depends on the minimizer of a convex optimization problem (denoted lower-level problem). We utilize an iterative algorithm called `piggyback' to compute the gradient of the loss and minimizer of the lower-level problem. Given that the lower-level problem is solved numerically, the loss function and thus its gradient can only be computed inexactly. To estimate the accuracy of the computed hypergradient, we derive an a-posteriori error bound, which provides guides for setting the tolerance for the lower-level problem, as well as the piggyback algorithm. To efficiently solve the upper-level optimization, we also propose an adaptive method for choosing a suitable step-size. To illustrate the proposed method, we consider a few learned regularizer problems, such as training an input-convex neural network.
☆ Integrating Expert Labels into LLM-based Emission Goal Detection: Example Selection vs Automatic Prompt Design
We address the detection of emission reduction goals in corporate reports, an important task for monitoring companies' progress in addressing climate change. Specifically, we focus on the issue of integrating expert feedback in the form of labeled example passages into LLM-based pipelines, and compare the two strategies of (1) a dynamic selection of few-shot examples and (2) the automatic optimization of the prompt by the LLM itself. Our findings on a public dataset of 769 climate-related passages from real-world business reports indicate that automatic prompt optimization is the superior approach, while combining both methods provides only limited benefit. Qualitative results indicate that optimized prompts do indeed capture many intricacies of the targeted emission goal extraction task.
☆ Federated Split Learning with Model Pruning and Gradient Quantization in Wireless Networks
As a paradigm of distributed machine learning, federated learning typically requires all edge devices to train a complete model locally. However, with the increasing scale of artificial intelligence models, the limited resources on edge devices often become a bottleneck for efficient fine-tuning. To address this challenge, federated split learning (FedSL) implements collaborative training across the edge devices and the server through model splitting. In this paper, we propose a lightweight FedSL scheme, that further alleviates the training burden on resource-constrained edge devices by pruning the client-side model dynamicly and using quantized gradient updates to reduce computation overhead. Additionally, we apply random dropout to the activation values at the split layer to reduce communication overhead. We conduct theoretical analysis to quantify the convergence performance of the proposed scheme. Finally, simulation results verify the effectiveness and advantages of the proposed lightweight FedSL in wireless network environments.
☆ BatchTopK Sparse Autoencoders
Sparse autoencoders (SAEs) have emerged as a powerful tool for interpreting language model activations by decomposing them into sparse, interpretable features. A popular approach is the TopK SAE, that uses a fixed number of the most active latents per sample to reconstruct the model activations. We introduce BatchTopK SAEs, a training method that improves upon TopK SAEs by relaxing the top-k constraint to the batch-level, allowing for a variable number of latents to be active per sample. As a result, BatchTopK adaptively allocates more or fewer latents depending on the sample, improving reconstruction without sacrificing average sparsity. We show that BatchTopK SAEs consistently outperform TopK SAEs in reconstructing activations from GPT-2 Small and Gemma 2 2B, and achieve comparable performance to state-of-the-art JumpReLU SAEs. However, an advantage of BatchTopK is that the average number of latents can be directly specified, rather than approximately tuned through a costly hyperparameter sweep. We provide code for training and evaluating BatchTopK SAEs at https://github.com/bartbussmann/BatchTopK
☆ Edge Delayed Deep Deterministic Policy Gradient: efficient continuous control for edge scenarios
Deep Reinforcement Learning is gaining increasing attention thanks to its capability to learn complex policies in high-dimensional settings. Recent advancements utilize a dual-network architecture to learn optimal policies through the Q-learning algorithm. However, this approach has notable drawbacks, such as an overestimation bias that can disrupt the learning process and degrade the performance of the resulting policy. To address this, novel algorithms have been developed that mitigate overestimation bias by employing multiple Q-functions. Edge scenarios, which prioritize privacy, have recently gained prominence. In these settings, limited computational resources pose a significant challenge for complex Machine Learning approaches, making the efficiency of algorithms crucial for their performance. In this work, we introduce a novel Reinforcement Learning algorithm tailored for edge scenarios, called Edge Delayed Deep Deterministic Policy Gradient (EdgeD3). EdgeD3 enhances the Deep Deterministic Policy Gradient (DDPG) algorithm, achieving significantly improved performance with $25\%$ less Graphics Process Unit (GPU) time while maintaining the same memory usage. Additionally, EdgeD3 consistently matches or surpasses the performance of state-of-the-art methods across various benchmarks, all while using $30\%$ fewer computational resources and requiring $30\%$ less memory.
☆ Exploring the Impact of Synthetic Data on Human Gesture Recognition Tasks Using GANs
In the evolving domain of Human Activity Recognition (HAR) using Internet of Things (IoT) devices, there is an emerging interest in employing Deep Generative Models (DGMs) to address data scarcity, enhance data quality, and improve classification metrics scores. Among these types of models, Generative Adversarial Networks (GANs) have arisen as a powerful tool for generating synthetic data that mimic real-world scenarios with high fidelity. However, Human Gesture Recognition (HGR), a subset of HAR, particularly in healthcare applications, using time series data such as allergic gestures, remains highly unexplored. In this paper, we examine and evaluate the performance of two GANs in the generation of synthetic gesture motion data that compose a part of an open-source benchmark dataset. The data is related to the disease identification domain and healthcare, specifically to allergic rhinitis. We also focus on these AI models' performance in terms of fidelity, diversity, and privacy. Furthermore, we examine the scenario if the synthetic data can substitute real data, in training scenarios and how well models trained on synthetic data can be generalized for the allergic rhinitis gestures. In our work, these gestures are related to 6-axes accelerometer and gyroscope data, serving as multi-variate time series instances, and retrieved from smart wearable devices. To the best of our knowledge, this study is the first to explore the feasibility of synthesizing motion gestures for allergic rhinitis from wearable IoT device data using Generative Adversarial Networks (GANs) and testing their impact on the generalization of gesture recognition systems. It is worth noting that, even if our method has been applied to a specific category of gestures, it is designed to be generalized and can be deployed also to other motion data in the HGR domain.
comment: 8 pages, 5 figures, 20th International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT), 2024
☆ PyPulse: A Python Library for Biosignal Imputation
We introduce PyPulse, a Python package for imputation of biosignals in both clinical and wearable sensor settings. Missingness is commonplace in these settings and can arise from multiple causes, such as insecure sensor attachment or data transmission loss. PyPulse's framework provides a modular and extendable framework with high ease-of-use for a broad userbase, including non-machine-learning bioresearchers. Specifically, its new capabilities include using pre-trained imputation methods out-of-the-box on custom datasets, running the full workflow of training or testing a baseline method with a single line of code, and comparing baseline methods in an interactive visualization tool. We released PyPulse under the MIT License on Github and PyPI. The source code can be found at: https://github.com/rehg-lab/pulseimpute.
comment: 7 pages, 3 figures. Implementation and documentation are available at https://github.com/rehg-lab/pulseimpute
☆ Gentle robustness implies Generalization
Robustness and generalization ability of machine learning models are of utmost importance in various application domains. There is a wide interest in efficient ways to analyze those properties. One important direction is to analyze connection between those two properties. Prior theories suggest that a robust learning algorithm can produce trained models with a high generalization ability. However, we show in this work that the existing error bounds are vacuous for the Bayes optimal classifier which is the best among all measurable classifiers for a classification problem with overlapping classes. Those bounds cannot converge to the true error of this ideal classifier. This is undesirable, surprizing, and never known before. We then present a class of novel bounds, which are model-dependent and provably tighter than the existing robustness-based ones. Unlike prior ones, our bounds are guaranteed to converge to the true error of the best classifier, as the number of samples increases. We further provide an extensive experiment and find that two of our bounds are often non-vacuous for a large class of deep neural networks, pretrained from ImageNet.
☆ Low-Rank Matrix Factorizations with Volume-based Constraints and Regularizations
Low-rank matrix factorizations are a class of linear models widely used in various fields such as machine learning, signal processing, and data analysis. These models approximate a matrix as the product of two smaller matrices, where the left matrix captures latent features while the right matrix linearly decomposes the data based on these features. There are many ways to define what makes a component "important." Standard LRMFs, such as the truncated singular value decomposition, focus on minimizing the distance between the original matrix and its low-rank approximation. In this thesis, the notion of "importance" is closely linked to interpretability and uniqueness, which are key to obtaining reliable and meaningful results. This thesis thus focuses on volume-based constraints and regularizations designed to enhance interpretability and uniqueness. We first introduce two new volume-constrained LRMFs designed to enhance these properties. The first assumes that data points are naturally bounded (e.g., movie ratings between 1 and 5 stars) and can be explained by convex combinations of features within the same bounds, allowing them to be interpreted in the same way as the data. The second model is more general, constraining the factors to belong to convex polytopes. Then, two variants of volume-regularized LRMFs are proposed. The first minimizes the volume of the latent features, encouraging them to cluster closely together, while the second maximizes the volume of the decompositions, promoting sparse representations. Across all these models, uniqueness is achieved under the core principle that the factors must be "sufficiently scattered" within their respective feasible sets. Motivated by applications such as blind source separation and missing data imputation, this thesis also proposes efficient algorithms that make these models practical for real-world applications.
☆ Exploring Memorization and Copyright Violation in Frontier LLMs: A Study of the New York Times v. OpenAI 2023 Lawsuit
Copyright infringement in frontier LLMs has received much attention recently due to the New York Times v. OpenAI lawsuit, filed in December 2023. The New York Times claims that GPT-4 has infringed its copyrights by reproducing articles for use in LLM training and by memorizing the inputs, thereby publicly displaying them in LLM outputs. Our work aims to measure the propensity of OpenAI's LLMs to exhibit verbatim memorization in its outputs relative to other LLMs, specifically focusing on news articles. We discover that both GPT and Claude models use refusal training and output filters to prevent verbatim output of the memorized articles. We apply a basic prompt template to bypass the refusal training and show that OpenAI models are currently less prone to memorization elicitation than models from Meta, Mistral, and Anthropic. We find that as models increase in size, especially beyond 100 billion parameters, they demonstrate significantly greater capacity for memorization. Our findings have practical implications for training: more attention must be placed on preventing verbatim memorization in very large models. Our findings also have legal significance: in assessing the relative memorization capacity of OpenAI's LLMs, we probe the strength of The New York Times's copyright infringement claims and OpenAI's legal defenses, while underscoring issues at the intersection of generative AI, law, and policy.
☆ Measuring Pre-training Data Quality without Labels for Time Series Foundation Models
Recently, there has been a growing interest in time series foundation models that generalize across different downstream tasks. A key to strong foundation models is a diverse pre-training dataset, which is particularly challenging to collect for time series classification. In this work, we explore the performance of a contrastive-learning-based foundation model as a function of the data used for pre-training. We introduce contrastive accuracy, a new measure to evaluate the quality of the representation space learned by the foundation model. Our experiments reveal the positive correlation between the proposed measure and the accuracy of the model on a collection of downstream tasks. This suggests that the contrastive accuracy can serve as a criterion to search for time series datasets that can enhance the pre-training and improve thereby the foundation model's generalization.
☆ GraphNeuralNetworks.jl: Deep Learning on Graphs with Julia
GraphNeuralNetworks.jl is an open-source framework for deep learning on graphs, written in the Julia programming language. It supports multiple GPU backends, generic sparse or dense graph representations, and offers convenient interfaces for manipulating standard, heterogeneous, and temporal graphs with attributes at the node, edge, and graph levels. The framework allows users to define custom graph convolutional layers using gather/scatter message-passing primitives or optimized fused operations. It also includes several popular layers, enabling efficient experimentation with complex deep architectures. The package is available on GitHub: \url{https://github.com/JuliaGraphs/GraphNeuralNetworks.jl}.
comment: Submitted to JMLR OSS
☆ Tracking control of latent dynamic systems with application to spacecraft attitude control
When intelligent spacecraft or space robots perform tasks in a complex environment, the controllable variables are usually not directly available and have to be inferred from high-dimensional observable variables, such as outputs of neural networks or images. While the dynamics of these observations are highly complex, the mechanisms behind them may be simple, which makes it possible to regard them as latent dynamic systems. For control of latent dynamic systems, methods based on reinforcement learning suffer from sample inefficiency and generalization problems. In this work, we propose an asymptotic tracking controller for latent dynamic systems. The latent variables are related to the high-dimensional observations through an unknown nonlinear function. The dynamics are unknown but assumed to be affine nonlinear. To realize asymptotic tracking, an identifiable latent dynamic model is learned to recover the latents and estimate the dynamics. This training process does not depend on the goals or reference trajectories. Based on the learned model, we use a manually designed feedback linearization controller to ensure the asymptotic tracking property of the closed-loop system. After considering fully controllable systems, the results are extended to the case that uncontrollable environmental latents exist. As an application, simulation experiments on a latent spacecraft attitude dynamic model are conducted to verify the proposed methods, and the observation noise and control deviation are taken into consideration.
☆ Augmenting the action space with conventions to improve multi-agent cooperation in Hanabi AAMAS
The card game Hanabi is considered a strong medium for the testing and development of multi-agent reinforcement learning (MARL) algorithms, due to its cooperative nature, hidden information, limited communication and remarkable complexity. Previous research efforts have explored the capabilities of MARL algorithms within Hanabi, focusing largely on advanced architecture design and algorithmic manipulations to achieve state-of-the-art performance for a various number of cooperators. However, this often leads to complex solution strategies with high computational cost and requiring large amounts of training data. For humans to solve the Hanabi game effectively, they require the use of conventions, which often allows for a means to implicitly convey ideas or knowledge based on a predefined, and mutually agreed upon, set of ``rules''. Multi-agent problems containing partial observability, especially when limited communication is present, can benefit greatly from the use of implicit knowledge sharing. In this paper, we propose a novel approach to augmenting the action space using conventions, which act as special cooperative actions that span over multiple time steps and multiple agents, requiring agents to actively opt in for it to reach fruition. These conventions are based on existing human conventions, and result in a significant improvement on the performance of existing techniques for self-play and cross-play across a various number of cooperators within Hanabi.
comment: This paper is under review at the journal of autonomous agents and multi-agent systems (JAAMAS)
☆ Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection SC
Automatic Speech Recognition (ASR) plays an important role in speech-based automatic detection of Alzheimer's disease (AD). However, recognition errors could propagate downstream, potentially impacting the detection decisions. Recent studies have revealed a non-linear relationship between word error rates (WER) and AD detection performance, where ASR transcriptions with notable errors could still yield AD detection accuracy equivalent to that based on manual transcriptions. This work presents a series of analyses to explore the effect of ASR transcription errors in BERT-based AD detection systems. Our investigation reveals that not all ASR errors contribute equally to detection performance. Certain words, such as stopwords, despite constituting a large proportion of errors, are shown to play a limited role in distinguishing AD. In contrast, the keywords related to diagnosis tasks exhibit significantly greater importance relative to other words. These findings provide insights into the interplay between ASR errors and the downstream detection model.
comment: Accepted by IEEE ISCSLP 2024
☆ Normalizing Flows are Capable Generative Models
Normalizing Flows (NFs) are likelihood-based models for continuous inputs. They have demonstrated promising results on both density estimation and generative modeling tasks, but have received relatively little attention in recent years. In this work, we demonstrate that NFs are more powerful than previously believed. We present \textit{TarFlow}: a simple and scalable architecture that enables highly performant NF models. TarFlow can be thought of as a Transformer-based variant of Masked Autoregressive Flows (MAFs): it consists of a stack of autoregressive Transformer blocks on image patches, alternating the autoregression direction between layers. TarFlow is straightforward to train end-to-end, and capable of directly modeling and generating pixels. We also propose three key techniques to improve sample quality: Gaussian noise augmentation during training, a post training denoising procedure, and an effective guidance method for both class-conditional and unconditional settings. Putting these together, TarFlow sets new state-of-the-art results on likelihood estimation for images, beating the previous best methods by a large margin, and generates samples with quality and diversity comparable to diffusion models, for the first time with a stand-alone NF model. We make our code available at \href{https://github.com/apple/ml-tarflow}{https://github.com/apple/ml-tarflow}.
☆ Vision-Based Deep Reinforcement Learning of UAV Autonomous Navigation Using Privileged Information
The capability of UAVs for efficient autonomous navigation and obstacle avoidance in complex and unknown environments is critical for applications in agricultural irrigation, disaster relief and logistics. In this paper, we propose the DPRL (Distributed Privileged Reinforcement Learning) navigation algorithm, an end-to-end policy designed to address the challenge of high-speed autonomous UAV navigation under partially observable environmental conditions. Our approach combines deep reinforcement learning with privileged learning to overcome the impact of observation data corruption caused by partial observability. We leverage an asymmetric Actor-Critic architecture to provide the agent with privileged information during training, which enhances the model's perceptual capabilities. Additionally, we present a multi-agent exploration strategy across diverse environments to accelerate experience collection, which in turn expedites model convergence. We conducted extensive simulations across various scenarios, benchmarking our DPRL algorithm against the state-of-the-art navigation algorithms. The results consistently demonstrate the superior performance of our algorithm in terms of flight efficiency, robustness and overall success rate.
comment: 19 pages, 9 figures
☆ DSAI: Unbiased and Interpretable Latent Feature Extraction for Data-Centric AI
Large language models (LLMs) often struggle to objectively identify latent characteristics in large datasets due to their reliance on pre-trained knowledge rather than actual data patterns. To address this data grounding issue, we propose Data Scientist AI (DSAI), a framework that enables unbiased and interpretable feature extraction through a multi-stage pipeline with quantifiable prominence metrics for evaluating extracted features. On synthetic datasets with known ground-truth features, DSAI demonstrates high recall in identifying expert-defined features while faithfully reflecting the underlying data. Applications on real-world datasets illustrate the framework's practical utility in uncovering meaningful patterns with minimal expert oversight, supporting use cases such as interpretable classification. The title of our paper is chosen from multiple candidates based on DSAI-generated criteria.
☆ S$^{2}$FT: Efficient, Scalable and Generalizable LLM Fine-tuning by Structured Sparsity
Current PEFT methods for LLMs can achieve either high quality, efficient training, or scalable serving, but not all three simultaneously. To address this limitation, we investigate sparse fine-tuning and observe a remarkable improvement in generalization ability. Utilizing this key insight, we propose a family of Structured Sparse Fine-Tuning (S$^{2}$FT) methods for LLMs, which concurrently achieve state-of-the-art fine-tuning performance, training efficiency, and inference scalability. S$^{2}$FT accomplishes this by "selecting sparsely and computing densely". It selects a few heads and channels in the MHA and FFN modules for each Transformer block, respectively. Next, it co-permutes weight matrices on both sides of the coupled structures in LLMs to connect the selected components in each layer into a dense submatrix. Finally, S$^{2}$FT performs in-place gradient updates on all submatrices. Through theoretical analysis and empirical results, our method prevents overfitting and forgetting, delivers SOTA performance on both commonsense and arithmetic reasoning with 4.6% and 1.3% average improvements compared to LoRA, and surpasses full FT by 11.5% when generalizing to various domains after instruction tuning. Using our partial backpropagation algorithm, S$^{2}$FT saves training memory up to 3$\times$ and improves latency by 1.5-2.7$\times$ compared to full FT, while delivering an average 10% improvement over LoRA on both metrics. We further demonstrate that the weight updates in S$^{2}$FT can be decoupled into adapters, enabling effective fusion, fast switch, and efficient parallelism for serving multiple fine-tuned models.
☆ Table2Image: Interpretable Tabular data Classification with Realistic Image Transformations
Recent advancements in deep learning for tabular data have demonstrated promising performance, yet interpretable models remain limited, with many relying on complex and large-scale architectures. This paper introduces Table2Image, an interpretable framework that transforms tabular data into realistic image representations for classification, achieving competitive performance with relatively lightweight models. Additionally, we propose variance inflation factor (VIF) initialization, which reflects the statistical properties of the data, and a novel interpretability framework that integrates insights from both the original tabular data and its image transformations. By leveraging Shapley additive explanations (SHAP) with methods to minimize distributional discrepancies, our approach combines tabular and image-based representations. Experiments on benchmark datasets showcase competitive classification accuracy, area under the curve (AUC), and improved interpretability, offering a scalable and reliable solution. Our code is available at https://github.com/duneag2/table2image.
☆ Flow Matching Guide and Code
Flow Matching (FM) is a recent framework for generative modeling that has achieved state-of-the-art performance across various domains, including image, video, audio, speech, and biological structures. This guide offers a comprehensive and self-contained review of FM, covering its mathematical foundations, design choices, and extensions. By also providing a PyTorch package featuring relevant examples (e.g., image and text generation), this work aims to serve as a resource for both novice and experienced researchers interested in understanding, applying and further developing FM.
☆ Optimizing Multi-Task Learning for Enhanced Performance in Large Language Models
This study aims to explore the performance improvement method of large language models based on GPT-4 under the multi-task learning framework and conducts experiments on two tasks: text classification and automatic summary generation. Through the combined design of shared feature extractors and task-specific modules, we achieve knowledge-sharing and optimization of multiple tasks in the same model. The experiment uses multiple subtasks of the GLUE dataset to compare the performance of the multi-task model with the single-task GPT-4, the multi-task version of GPT-3, the BERT basic model, and the classic Bi-LSTM with Attention model. The results show that the proposed multi-task learning model outperforms other comparison models in terms of text classification accuracy and ROUGE value of summary generation, demonstrating the advantages of multi-task learning in improving model generalization ability and collaborative learning between tasks. The model maintains a stable loss convergence rate during training, showing good learning efficiency and adaptability to the test set. This study verifies the applicability of the multi-task learning framework in large language models, especially in improving the model's ability to balance different tasks. In the future, with the combination of large language models and multimodal data and the application of dynamic task adjustment technology, the framework based on multi-task learning is expected to play a greater role in practical applications across fields and provide new ideas for the development of general artificial intelligence.
☆ In Silico Pharmacokinetic and Molecular Docking Studies of Natural Plants against Essential Protein KRAS for Treatment of Pancreatic Cancer
A kind of pancreatic cancer called Pancreatic Ductal Adenocarcinoma (PDAC) is anticipated to be one of the main causes of mortality during past years. Evidence from several researches supported the concept that the oncogenic KRAS (Ki-ras2 Kirsten rat sarcoma viral oncogene) mutation is the major cause of pancreatic cancer. KRAS acts as an on-off switch that promotes cell growth. But when the KRAS gene is mutated, it will be in one position, allowing the cell growth uncontrollably. This uncontrollable multiplication of cells causes cancer growth. Therefore, KRAS was selected as the target protein in the study. Fifty plant-derived compounds are selected for the study. To determine whether the examined drugs could bind to the KRAS complex's binding pocket, molecular docking was performed. Computational analyses were used to assess the possible ability of tested substances to pass the Blood Brain Barrier (BBB). To predict the bioactivity of ligands a machine learning model was created. Five machine learning models were created and have chosen the best one among them for analyzing the bioactivity of each ligand. From the fifty plant-derived compounds the compounds with the least binding energies are selected. Then bioactivity of these six compounds is analyzed using Random Forest Regression model. Adsorption, Distribution, Metabolism, Excretion (ADME) properties of compounds are analyzed. The results showed that borneol has powerful effects and acts as a promising agent for the treatment of pancreatic cancer. This suggests that borneol found in plants like mint, ginger, rosemary, etc., is a successful compound for the treatment of pancreatic cancer.
☆ VariFace: Fair and Diverse Synthetic Dataset Generation for Face Recognition
The use of large-scale, web-scraped datasets to train face recognition models has raised significant privacy and bias concerns. Synthetic methods mitigate these concerns and provide scalable and controllable face generation to enable fair and accurate face recognition. However, existing synthetic datasets display limited intraclass and interclass diversity and do not match the face recognition performance obtained using real datasets. Here, we propose VariFace, a two-stage diffusion-based pipeline to create fair and diverse synthetic face datasets to train face recognition models. Specifically, we introduce three methods: Face Recognition Consistency to refine demographic labels, Face Vendi Score Guidance to improve interclass diversity, and Divergence Score Conditioning to balance the identity preservation-intraclass diversity trade-off. When constrained to the same dataset size, VariFace considerably outperforms previous synthetic datasets (0.9200 $\rightarrow$ 0.9405) and achieves comparable performance to face recognition models trained with real data (Real Gap = -0.0065). In an unconstrained setting, VariFace not only consistently achieves better performance compared to previous synthetic methods across dataset sizes but also, for the first time, outperforms the real dataset (CASIA-WebFace) across six evaluation datasets. This sets a new state-of-the-art performance with an average face verification accuracy of 0.9567 (Real Gap = +0.0097) across LFW, CFP-FP, CPLFW, AgeDB, and CALFW datasets and 0.9366 (Real Gap = +0.0380) on the RFW dataset.
☆ Representational Transfer Learning for Matrix Completion
We propose to transfer representational knowledge from multiple sources to a target noisy matrix completion task by aggregating singular subspaces information. Under our representational similarity framework, we first integrate linear representation information by solving a two-way principal component analysis problem based on a properly debiased matrix-valued dataset. After acquiring better column and row representation estimators from the sources, the original high-dimensional target matrix completion problem is then transformed into a low-dimensional linear regression, of which the statistical efficiency is guaranteed. A variety of extensional arguments, including post-transfer statistical inference and robustness against negative transfer, are also discussed alongside. Finally, extensive simulation results and a number of real data cases are reported to support our claims.
☆ A Scalable Decentralized Reinforcement Learning Framework for UAV Target Localization Using Recurrent PPO
The rapid advancements in unmanned aerial vehicles (UAVs) have unlocked numerous applications, including environmental monitoring, disaster response, and agricultural surveying. Enhancing the collective behavior of multiple decentralized UAVs can significantly improve these applications through more efficient and coordinated operations. In this study, we explore a Recurrent PPO model for target localization in perceptually degraded environments like places without GNSS/GPS signals. We first developed a single-drone approach for target identification, followed by a decentralized two-drone model. Our approach can utilize two types of sensors on the UAVs, a detection sensor and a target signal sensor. The single-drone model achieved an accuracy of 93%, while the two-drone model achieved an accuracy of 86%, with the latter requiring fewer average steps to locate the target. This demonstrates the potential of our method in UAV swarms, offering efficient and effective localization of radiant targets in complex environmental conditions.
comment: Submitted to TENCON 2024
☆ A Self-guided Multimodal Approach to Enhancing Graph Representation Learning for Alzheimer's Diseases
Graph neural networks (GNNs) are powerful machine learning models designed to handle irregularly structured data. However, their generic design often proves inadequate for analyzing brain connectomes in Alzheimer's Disease (AD), highlighting the need to incorporate domain knowledge for optimal performance. Infusing AD-related knowledge into GNNs is a complicated task. Existing methods typically rely on collaboration between computer scientists and domain experts, which can be both time-intensive and resource-demanding. To address these limitations, this paper presents a novel self-guided, knowledge-infused multimodal GNN that autonomously incorporates domain knowledge into the model development process. Our approach conceptualizes domain knowledge as natural language and introduces a specialized multimodal GNN capable of leveraging this uncurated knowledge to guide the learning process of the GNN, such that it can improve the model performance and strengthen the interpretability of the predictions. To evaluate our framework, we curated a comprehensive dataset of recent peer-reviewed papers on AD and integrated it with multiple real-world AD datasets. Experimental results demonstrate the ability of our method to extract relevant domain knowledge, provide graph-based explanations for AD diagnosis, and improve the overall performance of the GNN. This approach provides a more scalable and efficient alternative to inject domain knowledge for AD compared with the manual design from the domain expert, advancing both prediction accuracy and interpretability in AD diagnosis.
☆ H-FedSN: Personalized Sparse Networks for Efficient and Accurate Hierarchical Federated Learning for IoT Applications
The proliferation of Internet of Things (IoT) has increased interest in federated learning (FL) for privacy-preserving distributed data utilization. However, traditional two-tier FL architectures inadequately adapt to multi-tier IoT environments. While Hierarchical Federated Learning (HFL) improves practicality in multi-tier IoT environments by multi-layer aggregation, it still faces challenges in communication efficiency and accuracy due to high data transfer volumes, data heterogeneity, and imbalanced device distribution, struggling to meet the low-latency and high-accuracy model training requirements of practical IoT scenarios. To overcome these limitations, we propose H-FedSN, an innovative approach for practical IoT environments. H-FedSN introduces a binary mask mechanism with shared and personalized layers to reduce communication overhead by creating a sparse network while keeping original weights frozen. To address data heterogeneity and imbalanced device distribution, we integrate personalized layers for local data adaptation and apply Bayesian aggregation with cumulative Beta distribution updates at edge and cloud levels, effectively balancing contributions from diverse client groups. Evaluations on three real-world IoT datasets and MNIST under non-IID settings demonstrate that H-FedSN significantly reduces communication costs by 58 to 238 times compared to HierFAVG while achieving high accuracy, making it highly effective for practical IoT applications in hierarchical federated learning scenarios.
☆ Skill-Enhanced Reinforcement Learning Acceleration from Demonstrations ICML 2024
Learning from Demonstration (LfD) aims to facilitate rapid Reinforcement Learning (RL) by leveraging expert demonstrations to pre-train the RL agent. However, the limited availability of expert demonstration data often hinders its ability to effectively aid downstream RL learning. To address this problem, we propose a novel two-stage method dubbed as Skill-enhanced Reinforcement Learning Acceleration (SeRLA). SeRLA introduces a skill-level adversarial Positive-Unlabeled (PU) learning model to extract useful skill prior knowledge by enabling learning from both limited expert data and general low-cost demonstration data in the offline prior learning stage. Subsequently, it deploys a skill-based soft actor-critic algorithm to leverage this acquired prior knowledge in the downstream online RL stage for efficient training of a skill policy network. Moreover, we develop a simple skill-level data enhancement technique to further alleviate data sparsity and improve both skill prior learning and downstream skill policy training. Our experimental results on multiple standard RL environments show the proposed SeRLA method achieves state-of-the-art performance on accelerating reinforcement learning on downstream tasks, especially in the early learning phase.
comment: ICML 2024 AutoRL Workshop; 9 pages
♻ ☆ Data Attribution for Text-to-Image Models by Unlearning Synthesized Images NeurIPS 2024
The goal of data attribution for text-to-image models is to identify the training images that most influence the generation of a new image. Influence is defined such that, for a given output, if a model is retrained from scratch without the most influential images, the model would fail to reproduce the same output. Unfortunately, directly searching for these influential images is computationally infeasible, since it would require repeatedly retraining models from scratch. In our work, we propose an efficient data attribution method by simulating unlearning the synthesized image. We achieve this by increasing the training loss on the output image, without catastrophic forgetting of other, unrelated concepts. We then identify training images with significant loss deviations after the unlearning process and label these as influential. We evaluate our method with a computationally intensive but "gold-standard" retraining from scratch and demonstrate our method's advantages over previous methods.
comment: Updated v2 -- NeurIPS 2024 camera ready version. Project page: https://peterwang512.github.io/AttributeByUnlearning Code: https://github.com/PeterWang512/AttributeByUnlearning
♻ ☆ SCADE: Scalable Framework for Anomaly Detection in High-Performance System SP
As command-line interfaces remain integral to high-performance computing environments, the risk of exploitation through stealthy and complex command-line abuse grows. Conventional security solutions struggle to detect these anomalies due to their context-specific nature, lack of labeled data, and the prevalence of sophisticated attacks like Living-off-the-Land (LOL). To address this gap, we introduce the Scalable Command-Line Anomaly Detection Engine (SCADE), a framework that combines global statistical models with local context-specific analysis for unsupervised anomaly detection. SCADE leverages novel statistical methods, including BM25 and Log Entropy, alongside dynamic thresholding to adaptively detect rare, malicious command-line patterns in low signal-to-noise ratio (SNR) environments. Experimental results show that SCADE achieves above 98% SNR in identifying anomalous behavior while minimizing false positives. Designed for scalability and precision, SCADE provides an innovative, metadata-enriched approach to anomaly detection, offering a robust solution for cybersecurity in high-computation environments. This work presents SCADE's architecture, detection methodology, and its potential for enhancing anomaly detection in enterprise systems. We argue that SCADE represents a significant advancement in unsupervised anomaly detection, offering a robust, adaptive framework for security analysts and researchers seeking to enhance detection accuracy in high-computation environments.
comment: Updated title and abstract for broader scope. Submitted to ACM CODASPY (The 15th ACM Conference on Data and Application Security and Privacy) Conference
♻ ☆ Croissant: A Metadata Format for ML-Ready Datasets NeurIPS 2024
Data is a critical resource for machine learning (ML), yet working with data remains a key friction point. This paper introduces Croissant, a metadata format for datasets that creates a shared representation across ML tools, frameworks, and platforms. Croissant makes datasets more discoverable, portable, and interoperable, thereby addressing significant challenges in ML data management. Croissant is already supported by several popular dataset repositories, spanning hundreds of thousands of datasets, enabling easy loading into the most commonly-used ML frameworks, regardless of where the data is stored. Our initial evaluation by human raters shows that Croissant metadata is readable, understandable, complete, yet concise.
comment: Published at the NeurIPS 2024 Datasets and Benchmark Track. A shorter version appeared earlier in Proceedings of ACM SIGMOD/PODS'24 Data Management for End-to-End Machine Learning (DEEM) Workshop https://dl.acm.org/doi/10.1145/3650203.3663326
♻ ☆ LLM Pruning and Distillation in Practice: The Minitron Approach
We present a comprehensive report on compressing the Llama 3.1 8B and Mistral NeMo 12B models to 4B and 8B parameters, respectively, using pruning and distillation. We explore two distinct pruning strategies: (1) depth pruning and (2) joint hidden/attention/MLP (width) pruning, and evaluate the results on common benchmarks from the LM Evaluation Harness. The models are then aligned with NeMo Aligner and tested in instruct-tuned versions. This approach produces a compelling 4B model from Llama 3.1 8B and a state-of-the-art Mistral-NeMo-Minitron-8B (MN-Minitron-8B for brevity) model from Mistral NeMo 12B. We found that with no access to the original data, it is beneficial to slightly fine-tune teacher models on the distillation dataset. We open-source our base model weights on Hugging Face with a permissive license.
comment: v4: Update author order
♻ ☆ The broader spectrum of in-context learning
The ability of language models to learn a task from a few examples in context has generated substantial interest. Here, we provide a perspective that situates this type of supervised few-shot learning within a much broader spectrum of meta-learned in-context learning. Indeed, we suggest that any distribution of sequences in which context non-trivially decreases loss on subsequent predictions can be interpreted as eliciting a kind of in-context learning. We suggest that this perspective helps to unify the broad set of in-context abilities that language models exhibit $\unicode{x2014}$ such as adapting to tasks from instructions or role play, or extrapolating time series. This perspective also sheds light on potential roots of in-context learning in lower-level processing of linguistic dependencies (e.g. coreference or parallel structures). Finally, taking this perspective highlights the importance of generalization, which we suggest can be studied along several dimensions: not only the ability to learn something novel, but also flexibility in learning from different presentations, and in applying what is learned. We discuss broader connections to past literature in meta-learning and goal-conditioned agents, and other perspectives on learning and adaptation. We close by suggesting that research on in-context learning should consider this broader spectrum of in-context capabilities and types of generalization.
♻ ☆ FabuLight-ASD: Unveiling Speech Activity via Body Language
Active speaker detection (ASD) in multimodal environments is crucial for various applications, from video conferencing to human-robot interaction. This paper introduces FabuLight-ASD, an advanced ASD model that integrates facial, audio, and body pose information to enhance detection accuracy and robustness. Our model builds upon the existing Light-ASD framework by incorporating human pose data, represented through skeleton graphs, which minimises computational overhead. Using the Wilder Active Speaker Detection (WASD) dataset, renowned for reliable face and body bounding box annotations, we demonstrate FabuLight-ASD's effectiveness in real-world scenarios. Achieving an overall mean average precision (mAP) of 94.3%, FabuLight-ASD outperforms Light-ASD, which has an overall mAP of 93.7% across various challenging scenarios. The incorporation of body pose information shows a particularly advantageous impact, with notable improvements in mAP observed in scenarios with speech impairment, face occlusion, and human voice background noise. Furthermore, efficiency analysis indicates only a modest increase in parameter count (27.3%) and multiply-accumulate operations (up to 2.4%), underscoring the model's efficiency and feasibility. These findings validate the efficacy of FabuLight-ASD in enhancing ASD performance through the integration of body pose data. FabuLight-ASD's code and model weights are available at https://github.com/knowledgetechnologyuhh/FabuLight-ASD.
comment: 23 pages, 8 figures, 3 tables, accepted for publication in Neural Computing and Applications
♻ ☆ FIARSE: Model-Heterogeneous Federated Learning via Importance-Aware Submodel Extraction NeurIPS 2024
In federated learning (FL), accommodating clients' varied computational capacities poses a challenge, often limiting the participation of those with constrained resources in global model training. To address this issue, the concept of model heterogeneity through submodel extraction has emerged, offering a tailored solution that aligns the model's complexity with each client's computational capacity. In this work, we propose Federated Importance-Aware Submodel Extraction (FIARSE), a novel approach that dynamically adjusts submodels based on the importance of model parameters, thereby overcoming the limitations of previous static and dynamic submodel extraction methods. Compared to existing works, the proposed method offers a theoretical foundation for the submodel extraction and eliminates the need for additional information beyond the model parameters themselves to determine parameter importance, significantly reducing the overhead on clients. Extensive experiments are conducted on various datasets to showcase the superior performance of the proposed FIARSE.
comment: This paper has been accepted by NeurIPS 2024. In this updated version, we have corrected the typos
♻ ☆ DexDiffuser: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulation
Dexterous manipulation with contact-rich interactions is crucial for advanced robotics. While recent diffusion-based planning approaches show promise for simpler manipulation tasks, they often produce unrealistic ghost states (e.g., the object automatically moves without hand contact) or lack adaptability when handling complex sequential interactions. In this work, we introduce DexDiffuser, an interaction-aware diffusion planning framework for adaptive dexterous manipulation. DexDiffuser models joint state-action dynamics through a dual-phase diffusion process which consists of pre-interaction contact alignment and post-contact goal-directed control, enabling goal-adaptive generalizable dexterous manipulation. Additionally, we incorporate dynamics model-based dual guidance and leverage large language models for automated guidance function generation, enhancing generalizability for physical interactions and facilitating diverse goal adaptation through language cues. Experiments on physical interaction tasks such as door opening, pen and block re-orientation, and hammer striking demonstrate DexDiffuser's effectiveness on goals outside training distributions, achieving over twice the average success rate (59.2% vs. 29.5%) compared to existing methods. Our framework achieves 70.0% success on 30-degree door opening, 40.0% and 36.7% on pen and block half-side re-orientation respectively, and 46.7% on hammer nail half drive, highlighting its robustness and flexibility in contact-rich manipulation.
comment: 27 pages (with new appendix). Project page: https://dexdiffuser.github.io/
♻ ☆ Distributed Thompson sampling under constrained communication
In Bayesian optimization, a black-box function is maximized via the use of a surrogate model. We apply distributed Thompson sampling, using a Gaussian process as a surrogate model, to approach the multi-agent Bayesian optimization problem. In our distributed Thompson sampling implementation, each agent receives sampled points from neighbors, where the communication network is encoded in a graph; each agent utilizes their own Gaussian process to model the objective function. We demonstrate theoretical bounds on Bayesian simple regret and Bayesian average regret, where the bound depends on the structure of the communication graph. Unlike in batch Bayesian optimization, this bound is applicable in cases where the communication graph amongst agents is constrained. When compared to sequential single-agent Thompson sampling, our bound guarantees faster convergence with respect to time as long as the communication graph is connected. We confirm the efficacy of our algorithm with numerical simulations on traditional optimization test functions, illustrating the significance of graph connectivity on improving regret convergence.
comment: 9 pages
♻ ☆ Optimistic Query Routing in Clustering-based Approximate Maximum Inner Product Search
Clustering-based nearest neighbor search is an effective method in which points are partitioned into geometric shards to form an index, with only a few shards searched during query processing to find a set of top-$k$ vectors. Even though the search efficacy is heavily influenced by the algorithm that identifies the shards to probe, it has received little attention in the literature. This work bridges that gap by studying routing in clustering-based maximum inner product search. We unpack existing routers and notice the surprising contribution of optimism. We then take a page from the sequential decision making literature and formalize that insight following the principle of ``optimism in the face of uncertainty.'' In particular, we present a framework that incorporates the moments of the distribution of inner products within each shard to estimate the maximum inner product. We then present an instance of our algorithm that uses only the first two moments to reach the same accuracy as state-of-the-art routers such as ScaNN by probing up to $50\%$ fewer points on benchmark datasets. Our algorithm is also space-efficient: we design a sketch of the second moment whose size is independent of the number of points and requires $\mathcal{O}(1)$ vectors per shard.
♻ ☆ Break a Lag: Triple Exponential Moving Average for Enhanced Optimization
The performance of deep learning models is critically dependent on sophisticated optimization strategies. While existing optimizers have shown promising results, many rely on first-order Exponential Moving Average (EMA) techniques, which often limit their ability to track complex gradient trends accurately. This fact can lead to a significant lag in trend identification and suboptimal optimization, particularly in highly dynamic gradient behavior. To address this fundamental limitation, we introduce Fast Adaptive Moment Estimation (FAME), a novel optimization technique that leverages the power of Triple Exponential Moving Average. By incorporating an advanced tracking mechanism, FAME enhances responsiveness to data dynamics, mitigates trend identification lag, and optimizes learning efficiency. Our comprehensive evaluation encompasses different computer vision tasks including image classification, object detection, and semantic segmentation, integrating FAME into 30 distinct architectures ranging from lightweight CNNs to Vision Transformers. Through rigorous benchmarking against state-of-the-art optimizers, FAME demonstrates superior accuracy and robustness. Notably, it offers high scalability, delivering substantial improvements across diverse model complexities, architectures, tasks, and benchmarks.
♻ ☆ How transformers learn structured data: insights from hierarchical filtering
Understanding the learning process and the embedded computation in transformers is becoming a central goal for the development of interpretable AI. In the present study, we introduce a hierarchical filtering procedure for generative models of sequences on trees, allowing us to hand-tune the range of positional correlations in the data. Leveraging this controlled setting, we provide evidence that vanilla encoder-only transformers can approximate the exact inference algorithm when trained on root classification and masked language modeling tasks, and study how this computation is discovered and implemented. We find that correlations at larger distances, corresponding to increasing layers of the hierarchy, are sequentially included by the network during training. Moreover, by comparing attention maps from models trained with varying degrees of filtering and by probing the different encoder levels, we find clear evidence of a reconstruction of correlations on successive length scales corresponding to the various levels of the hierarchy, which we relate to a plausible implementation of the exact inference algorithm within the same architecture.
comment: 21 pages, 15 figures
♻ ☆ Can tweets predict article retractions? A comparison between human and LLM labelling
Quickly detecting problematic research articles is crucial to safeguarding the integrity of scientific research. This study explores whether Twitter mentions of retracted articles can signal potential problems with the articles prior to their retraction, potentially serving as an early warning system for scholars. To investigate this, we analysed a dataset of 4,354 Twitter mentions associated with 504 retracted articles. The effectiveness of Twitter mentions in predicting article retractions was evaluated by both manual and Large Language Model (LLM) labelling. Manual labelling results indicated that 25.7% of tweets signalled problems before retraction. Using the manual labelling results as the baseline, we found that LLMs (GPT-4o-mini, Gemini 1.5 Flash, and Claude-3.5-Haiku) outperformed lexicon-based sentiment analysis tools (e.g., TextBlob) in detecting potential problems, suggesting that automatic detection of problematic articles from social media using LLMs is technically feasible. Nevertheless, since only a small proportion of retracted articles (11.1%) were criticised on Twitter prior to retraction, such automatic systems would detect only a minority of problematic articles. Overall, this study offers insights into how social media data, coupled with emerging generative AI techniques, can support research integrity.
comment: 19 pages
♻ ☆ Extraction Propagation
Running backpropagation end to end on large neural networks is fraught with difficulties like vanishing gradients and degradation. In this paper we present an alternative architecture composed of many small neural networks that interact with one another. Instead of propagating gradients back through the architecture we propagate vector-valued messages computed via forward passes, which are then used to update the parameters. Currently the performance is conjectured as we are yet to implement the architecture. However, we do back it up with some theory. A previous version of this paper was entitled "Fusion encoder networks" and detailed a slightly different architecture.
♻ ☆ SigKAN: Signature-Weighted Kolmogorov-Arnold Networks for Time Series
We propose a novel approach that enhances multivariate function approximation using learnable path signatures and Kolmogorov-Arnold networks (KANs). We enhance the learning capabilities of these networks by weighting the values obtained by KANs using learnable path signatures, which capture important geometric features of paths. This combination allows for a more comprehensive and flexible representation of sequential and temporal data. We demonstrate through studies that our SigKANs with learnable path signatures perform better than conventional methods across a range of function approximation challenges. By leveraging path signatures in neural networks, this method offers intriguing opportunities to enhance performance in time series analysis and time series forecasting, among other fields.
comment: arXiv admin note: text overlap with arXiv:2405.07344, arXiv:2406.02486
♻ ☆ How Transformers Solve Propositional Logic Problems: A Mechanistic Analysis
Large language models (LLMs) have shown amazing performance on tasks that require planning and reasoning. Motivated by this, we investigate the internal mechanisms that underpin a network's ability to perform complex logical reasoning. We first construct a synthetic propositional logic problem that serves as a concrete test-bed for network training and evaluation. Crucially, this problem demands nontrivial planning to solve. We perform our study on two fronts. First, we pursue an understanding of precisely how a three-layer transformer, trained from scratch and attains perfect test accuracy, solves this problem. We are able to identify certain "planning" and "reasoning" mechanisms in the network that necessitate cooperation between the attention blocks to implement the desired logic. Second, we study how pretrained LLMs, namely Mistral-7B and Gemma-2-9B, solve this problem. We characterize their reasoning circuits through causal intervention experiments, providing necessity and sufficiency evidence for the circuits. We find evidence suggesting that the two models' latent reasoning strategies are surprisingly similar, and human-like. Overall, our work systemically uncovers novel aspects of small and large transformers, and continues the study of how they plan and reason.
♻ ☆ Code-as-Monitor: Constraint-aware Visual Programming for Reactive and Proactive Robotic Failure Detection
Automatic detection and prevention of open-set failures are crucial in closed-loop robotic systems. Recent studies often struggle to simultaneously identify unexpected failures reactively after they occur and prevent foreseeable ones proactively. To this end, we propose Code-as-Monitor (CaM), a novel paradigm leveraging the vision-language model (VLM) for both open-set reactive and proactive failure detection. The core of our method is to formulate both tasks as a unified set of spatio-temporal constraint satisfaction problems and use VLM-generated code to evaluate them for real-time monitoring. To enhance the accuracy and efficiency of monitoring, we further introduce constraint elements that abstract constraint-related entities or their parts into compact geometric elements. This approach offers greater generality, simplifies tracking, and facilitates constraint-aware visual programming by leveraging these elements as visual prompts. Experiments show that CaM achieves a 28.7% higher success rate and reduces execution time by 31.8% under severe disturbances compared to baselines across three simulators and a real-world setting. Moreover, CaM can be integrated with open-loop control policies to form closed-loop systems, enabling long-horizon tasks in cluttered scenes with dynamic environments.
comment: Project page: https://zhoues.github.io/Code-as-Monitor/
♻ ☆ APOLLO: SGD-like Memory, AdamW-level Performance
Large language models (LLMs) are notoriously memory-intensive during training, particularly with the popular AdamW optimizer. This memory burden necessitates using more or higher-end GPUs or reducing batch sizes, limiting training scalability and throughput. To address this, various memory-efficient optimizers have been proposed to reduce optimizer memory usage. However, they face critical challenges: (i) reliance on costly SVD operations; (ii) significant performance trade-offs compared to AdamW; and (iii) still substantial optimizer memory overhead to maintain competitive performance. In this work, we identify that AdamW's learning rate adaptation rule can be effectively coarsened as a structured learning rate update. Based on this insight, we propose Approximated Gradient Scaling for Memory-Efficient LLM Optimization (APOLLO), which approximates learning rate scaling using an auxiliary low-rank optimizer state based on pure random projection. This structured learning rate update rule makes APOLLO highly tolerant to further memory reductions while delivering comparable pre-training performance. Even its rank-1 variant, APOLLO-Mini, achieves superior pre-training performance compared to AdamW with SGD-level memory costs. Extensive experiments demonstrate that the APOLLO series performs on-par with or better than AdamW, while achieving greater memory savings by nearly eliminating the optimization states of AdamW. These savings provide significant system-level benefits: (1) Enhanced Throughput: 3x throughput on an 8xA100-80GB setup compared to AdamW by supporting 4x larger batch sizes. (2) Improved Model Scalability: Pre-training LLaMA-13B with naive DDP on A100-80GB GPUs without system-level optimizations. (3) Low-End GPU Friendly Pre-training: Pre-training LLaMA-7B on a single GPU using less than 12 GB of memory with weight quantization.
comment: Preprint
♻ ☆ Enhancing predictive imaging biomarker discovery through treatment effect analysis WACV 2025
Identifying predictive covariates, which forecast individual treatment effectiveness, is crucial for decision-making across different disciplines such as personalized medicine. These covariates, referred to as biomarkers, are extracted from pre-treatment data, often within randomized controlled trials, and should be distinguished from prognostic biomarkers, which are independent of treatment assignment. Our study focuses on discovering predictive imaging biomarkers, specific image features, by leveraging pre-treatment images to uncover new causal relationships. Unlike labor-intensive approaches relying on handcrafted features prone to bias, we present a novel task of directly learning predictive features from images. We propose an evaluation protocol to assess a model's ability to identify predictive imaging biomarkers and differentiate them from purely prognostic ones by employing statistical testing and a comprehensive analysis of image feature attribution. We explore the suitability of deep learning models originally developed for estimating the conditional average treatment effect (CATE) for this task, which have been assessed primarily for their precision of CATE estimation while overlooking the evaluation of imaging biomarker discovery. Our proof-of-concept analysis demonstrates the feasibility and potential of our approach in discovering and validating predictive imaging biomarkers from synthetic outcomes and real-world image datasets. Our code is available at \url{https://github.com/MIC-DKFZ/predictive_image_biomarker_analysis}.
comment: Accepted to WACV 2025
♻ ☆ Synergistic Development of Perovskite Memristors and Algorithms for Robust Analog Computing
Analog computing using non-volatile memristors has emerged as a promising solution for energy-efficient deep learning. New materials, like perovskites-based memristors are recently attractive due to their cost-effectiveness, energy efficiency and flexibility. Yet, challenges in material diversity and immature fabrications require extensive experimentation for device development. Moreover, significant non-idealities in these memristors often impede them for computing. Here, we propose a synergistic methodology to concurrently optimize perovskite memristor fabrication and develop robust analog DNNs that effectively address the inherent non-idealities of these memristors. Employing Bayesian optimization (BO) with a focus on usability, we efficiently identify optimal materials and fabrication conditions for perovskite memristors. Meanwhile, we developed "BayesMulti", a DNN training strategy utilizing BO-guided noise injection to improve the resistance of analog DNNs to memristor imperfections. Our approach theoretically ensures that within a certain range of parameter perturbations due to memristor non-idealities, the prediction outcomes remain consistent. Our integrated approach enables use of analog computing in much deeper and wider networks, which significantly outperforms existing methods in diverse tasks like image classification, autonomous driving, species identification, and large vision-language models, achieving up to 100-fold improvements. We further validate our methodology on a 10$\times$10 optimized perovskite memristor crossbar, demonstrating high accuracy in a classification task and low energy consumption. This study offers a versatile solution for efficient optimization of various analog computing systems, encompassing both devices and algorithms.
♻ ☆ Surpassing Cosine Similarity for Multidimensional Comparisons: Dimension Insensitive Euclidean Metric (DIEM)
Advancements in computational power and hardware efficiency have enabled the tackling of increasingly complex and high-dimensional problems. While artificial intelligence (AI) achieved remarkable results, the interpretability of high-dimensional solutions remains challenging. A critical issue is the comparison of multidimensional quantities, which is essential in techniques like Principal Component Analysis (PCA), or k-means clustering. Common metrics such as cosine similarity, Euclidean distance, and Manhattan distance are often used for such comparisons - for example in muscular synergies of the human motor control system. However, their applicability and interpretability diminish as dimensionality increases. This paper provides a comprehensive analysis of the effects of dimensionality on these metrics. Our results reveal significant limitations of cosine similarity, particularly its dependency on the dimensionality of the vectors, leading to biased and poorly interpretable outcomes. To address this, we introduce the Dimension Insensitive Euclidean Metric (DIEM) which demonstrates superior robustness and generalizability across dimensions. DIEM maintains consistent variability and eliminates the biases observed in traditional metrics, making it a reliable tool for high-dimensional comparisons. This novel metric has the potential to replace cosine similarity, providing a more accurate and insightful method to analyze multidimensional data in fields ranging from neuromotor control to machine and deep learning.
comment: 13 pages, 17 figures
♻ ☆ Deep learning for predicting the occurrence of tipping points
Tipping points occur in many real-world systems, at which the system shifts suddenly from one state to another. The ability to predict the occurrence of tipping points from time series data remains an outstanding challenge and a major interest in a broad range of research fields. Particularly, the widely used methods based on bifurcation theory are neither reliable in prediction accuracy nor applicable for irregularly-sampled time series which are commonly observed from real-world systems. Here we address this challenge by developing a deep learning algorithm for predicting the occurrence of tipping points in untrained systems, by exploiting information about normal forms. Our algorithm not only outperforms traditional methods for regularly-sampled model time series but also achieves accurate predictions for irregularly-sampled model time series and empirical time series. Our ability to predict tipping points for complex systems paves the way for mitigation risks, prevention of catastrophic failures, and restoration of degraded systems, with broad applications in social science, engineering, and biology.
♻ ☆ What If We Had Used a Different App? Reliable Counterfactual KPI Analysis in Wireless Systems
In modern wireless network architectures, such as Open Radio Access Network (O-RAN), the operation of the radio access network (RAN) is managed by applications, or apps for short, deployed at intelligent controllers. These apps are selected from a given catalog based on current contextual information. For instance, a scheduling app may be selected on the basis of current traffic and network conditions. Once an app is chosen and run, it is no longer possible to directly test the key performance indicators (KPIs) that would have been obtained with another app. In other words, we can never simultaneously observe both the actual KPI, obtained by the selected app, and the counterfactual KPI, which would have been attained with another app, for the same network condition, making individual-level counterfactual KPIs analysis particularly challenging. This what-if analysis, however, would be valuable to monitor and optimize the network operation, e.g., to identify suboptimal app selection strategies. This paper addresses the problem of estimating the values of KPIs that would have been obtained if a different app had been implemented by the RAN. To this end, we propose a conformal-prediction-based counterfactual analysis method for wireless systems that provides reliable error bars for the estimated KPIs, despite the inherent covariate shift between logged and test data. Experimental results for medium access control-layer apps and for physical-layer apps demonstrate the merits of the proposed method.
comment: This paper has been submitted to a journal
♻ ☆ Neural Network Surrogate and Projected Gradient Descent for Fast and Reliable Finite Element Model Calibration: a Case Study on an Intervertebral Disc
Accurate calibration of finite element (FE) models is essential across various biomechanical applications, including human intervertebral discs (IVDs), to ensure their reliability and use in diagnosing and planning treatments. However, traditional calibration methods are computationally intensive, requiring iterative, derivative-free optimization algorithms that often take days to converge. This study addresses these challenges by introducing a novel, efficient, and effective calibration method demonstrated on a human L4-L5 IVD FE model as a case study using a neural network (NN) surrogate. The NN surrogate predicts simulation outcomes with high accuracy, outperforming other machine learning models, and significantly reduces the computational cost associated with traditional FE simulations. Next, a Projected Gradient Descent (PGD) approach guided by gradients of the NN surrogate is proposed to efficiently calibrate FE models. Our method explicitly enforces feasibility with a projection step, thus maintaining material bounds throughout the optimization process. The proposed method is evaluated against SOTA Genetic Algorithm and inverse model baselines on synthetic and in vitro experimental datasets. Our approach demonstrates superior performance on synthetic data, achieving an MAE of 0.06 compared to the baselines' MAE of 0.18 and 0.54, respectively. On experimental specimens, our method outperforms the baseline in 5 out of 6 cases. While our approach requires initial dataset generation and surrogate training, these steps are performed only once, and the actual calibration takes under three seconds. In contrast, traditional calibration time scales linearly with the number of specimens, taking up to 8 days in the worst-case. Such efficiency paves the way for applying more complex FE models, potentially extending beyond IVDs, and enabling accurate patient-specific simulations.
comment: In review. Project code: https://github.com/matanat/IVD-CalibNN/
♻ ☆ Burning RED: Unlocking Subtask-Driven Reinforcement Learning and Risk-Awareness in Average-Reward Markov Decision Processes
Average-reward Markov decision processes (MDPs) provide a foundational framework for sequential decision-making under uncertainty. However, average-reward MDPs have remained largely unexplored in reinforcement learning (RL) settings, with the majority of RL-based efforts having been allocated to episodic and discounted MDPs. In this work, we study a unique structural property of average-reward MDPs and utilize it to introduce Reward-Extended Differential (or RED) reinforcement learning: a novel RL framework that can be used to effectively and efficiently solve various subtasks simultaneously in the average-reward setting. We introduce a family of RED learning algorithms for prediction and control, including proven-convergent algorithms for the tabular case. We then showcase the power of these algorithms by demonstrating how they can be used to learn a policy that optimizes, for the first time, the well-known conditional value-at-risk (CVaR) risk measure in a fully-online manner, without the use of an explicit bi-level optimization scheme or an augmented state-space.
♻ ☆ A Cognac shot to forget bad memories: Corrective Unlearning in GNNs
Graph Neural Networks (GNNs) are increasingly being used for a variety of ML applications on graph data. Because graph data does not follow the independently and identically distributed (i.i.d.) assumption, adversarial manipulations or incorrect data can propagate to other data points through message passing, which deteriorates the model's performance. To allow model developers to remove the adverse effects of manipulated entities from a trained GNN, we study the recently formulated problem of Corrective Unlearning. We find that current graph unlearning methods fail to unlearn the effect of manipulations even when the whole manipulated set is known. We introduce a new graph unlearning method, Cognac, which can unlearn the effect of the manipulation set even when only 5% of it is identified. It recovers most of the performance of a strong oracle with fully corrected training data, even beating retraining from scratch without the deletion set while being 8x more efficient. We hope our work assists GNN developers in mitigating harmful effects caused by issues in real-world data post-training. Our code is publicly available at https://github.com/varshitakolipaka/corrective-unlearning-for-gnns
♻ ☆ Similarity Equivariant Graph Neural Networks for Homogenization of Metamaterials
Soft, porous mechanical metamaterials exhibit pattern transformations that may have important applications in soft robotics, sound reduction and biomedicine. To design these innovative materials, it is important to be able to simulate them accurately and quickly, in order to tune their mechanical properties. Since conventional simulations using the finite element method entail a high computational cost, in this article we aim to develop a machine learning-based approach that scales favorably to serve as a surrogate model. To ensure that the model is also able to handle various microstructures, including those not encountered during training, we include the microstructure as part of the network input. Therefore, we introduce a graph neural network that predicts global quantities (energy, stress stiffness) as well as the pattern transformations that occur (the kinematics). To make our model as accurate and data-efficient as possible, various symmetries are incorporated into the model. The starting point is an E(n)-equivariant graph neural network (which respects translation, rotation and reflection) that has periodic boundary conditions (i.e., it is in-/equivariant with respect to the choice of RVE), is scale in-/equivariant, can simulate large deformations, and can predict scalars, vectors as well as second and fourth order tensors (specifically energy, stress and stiffness). The incorporation of scale equivariance makes the model equivariant with respect to the similarities group, of which the Euclidean group E(n) is a subgroup. We show that this network is more accurate and data-efficient than graph neural networks with fewer symmetries. To create an efficient graph representation of the finite element discretization, we use only the internal geometrical hole boundaries from the finite element mesh to achieve a better speed-up and scaling with the mesh size.
comment: 60 pages, 22 figures. Submitted to CMAME (Computer Methods in Applied Mechanics and Engineering)
♻ ☆ Implicit Dynamical Flow Fusion (IDFF) for Generative Modeling
Conditional Flow Matching (CFM) models can generate high-quality samples from a non-informative prior, but they can be slow, often needing hundreds of network evaluations (NFE). To address this, we propose Implicit Dynamical Flow Fusion (IDFF); IDFF learns a new vector field with an additional momentum term that enables taking longer steps during sample generation while maintaining the fidelity of the generated distribution. Consequently, IDFFs reduce the NFEs by a factor of ten (relative to CFMs) without sacrificing sample quality, enabling rapid sampling and efficient handling of image and time-series data generation tasks. We evaluate IDFF on standard benchmarks such as CIFAR-10 and CelebA for image generation, where we achieve likelihood and quality performance comparable to CFMs and diffusion-based models with fewer NFEs. IDFF also shows superior performance on time-series datasets modeling, including molecular simulation and sea surface temperature (SST) datasets, highlighting its versatility and effectiveness across different domains.\href{https://github.com/MrRezaeiUofT/IDFF}{Github Repository}
♻ ☆ Data Augmentation for Seizure Prediction with Generative Diffusion Model
Data augmentation (DA) can significantly strengthen the electroencephalogram (EEG)-based seizure prediction methods. However, existing DA approaches are just the linear transformations of original data and cannot explore the feature space to increase diversity effectively. Therefore, we propose a novel diffusion-based DA method called DiffEEG. DiffEEG can fully explore data distribution and generate samples with high diversity, offering extra information to classifiers. It involves two processes: the diffusion process and the denoised process. In the diffusion process, the model incrementally adds noise with different scales to EEG input and converts it into random noise. In this way, the representation of data can be learned. In the denoised process, the model utilizes learned knowledge to sample synthetic data from random noise input by gradually removing noise. The randomness of input noise and the precise representation enable the synthetic samples to possess diversity while ensuring the consistency of feature space. We compared DiffEEG with original, down-sampling, sliding windows and recombination methods, and integrated them into five representative classifiers. The experiments demonstrate the effectiveness and generality of our method. With the contribution of DiffEEG, the Multi-scale CNN achieves state-of-the-art performance, with an average sensitivity, FPR, AUC of 95.4%, 0.051/h, 0.932 on the CHB-MIT database and 93.6%, 0.121/h, 0.822 on the Kaggle database.
comment: 15 pages, 9 figures
♻ ☆ Efficient Task Grouping Through Samplewise Optimisation Landscape Analysis
Shared training approaches, such as multi-task learning (MTL) and gradient-based meta-learning, are widely used in various machine learning applications, but they often suffer from negative transfer, leading to performance degradation in specific tasks. While several optimisation techniques have been developed to mitigate this issue for pre-selected task cohorts, identifying optimal task combinations for joint learning - known as task grouping - remains underexplored and computationally challenging due to the exponential growth in task combinations and the need for extensive training and evaluation cycles. This paper introduces an efficient task grouping framework designed to reduce these overwhelming computational demands of the existing methods. The proposed framework infers pairwise task similarities through a sample-wise optimisation landscape analysis, eliminating the need for the shared model training required to infer task similarities in existing methods. With task similarities acquired, a graph-based clustering algorithm is employed to pinpoint near-optimal task groups, providing an approximate yet efficient and effective solution to the originally NP-hard problem. Empirical assessments conducted on 8 different datasets highlight the effectiveness of the proposed framework, revealing a five-fold speed enhancement compared to previous state-of-the-art methods. Moreover, the framework consistently demonstrates comparable performance, confirming its remarkable efficiency and effectiveness in task grouping.
comment: Under review at IEEE Transactions on Pattern Analysis and Machine Intelligence
♻ ☆ On the Convergence of Gradient Descent for Large Learning Rates
A vast literature on convergence guarantees for gradient descent and derived methods exists at the moment. However, a simple practical situation remains unexplored: when a fixed step size is used, can we expect gradient descent to converge starting from any initialization? We provide fundamental impossibility results showing that convergence becomes impossible no matter the initialization if the step size gets too big. Looking at the asymptotic value of the gradient norm along the optimization trajectory, we see that there is a sharp transition as the step size crosses a critical value. This has been observed by practitioners, yet the true mechanisms through which this happens remain unclear beyond heuristics. Using results from dynamical systems theory, we provide a proof of this in the case of linear neural networks with a squared loss. We also prove the impossibility of convergence for more general losses without requiring strong assumptions such as Lipschitz continuity for the gradient. We validate our findings through experiments with non-linear networks.
♻ ☆ From Pixels to Words: Leveraging Explainability in Face Recognition through Interactive Natural Language Processing
Face Recognition (FR) has advanced significantly with the development of deep learning, achieving high accuracy in several applications. However, the lack of interpretability of these systems raises concerns about their accountability, fairness, and reliability. In the present study, we propose an interactive framework to enhance the explainability of FR models by combining model-agnostic Explainable Artificial Intelligence (XAI) and Natural Language Processing (NLP) techniques. The proposed framework is able to accurately answer various questions of the user through an interactive chatbot. In particular, the explanations generated by our proposed method are in the form of natural language text and visual representations, which for example can describe how different facial regions contribute to the similarity measure between two faces. This is achieved through the automatic analysis of the output's saliency heatmaps of the face images and a BERT question-answering model, providing users with an interface that facilitates a comprehensive understanding of the FR decisions. The proposed approach is interactive, allowing the users to ask questions to get more precise information based on the user's background knowledge. More importantly, in contrast to previous studies, our solution does not decrease the face recognition performance. We demonstrate the effectiveness of the method through different experiments, highlighting its potential to make FR systems more interpretable and user-friendly, especially in sensitive applications where decision-making transparency is crucial.
♻ ☆ Imitating Language via Scalable Inverse Reinforcement Learning NeurIPS 2024
The majority of language model training builds on imitation learning. It covers pretraining, supervised fine-tuning, and affects the starting conditions for reinforcement learning from human feedback (RLHF). The simplicity and scalability of maximum likelihood estimation (MLE) for next token prediction led to its role as predominant paradigm. However, the broader field of imitation learning can more effectively utilize the sequential structure underlying autoregressive generation. We focus on investigating the inverse reinforcement learning (IRL) perspective to imitation, extracting rewards and directly optimizing sequences instead of individual token likelihoods and evaluate its benefits for fine-tuning large language models. We provide a new angle, reformulating inverse soft-Q-learning as a temporal difference regularized extension of MLE. This creates a principled connection between MLE and IRL and allows trading off added complexity with increased performance and diversity of generations in the supervised fine-tuning (SFT) setting. We find clear advantages for IRL-based imitation, in particular for retaining diversity while maximizing task performance, rendering IRL a strong alternative on fixed SFT datasets even without online data generation. Our analysis of IRL-extracted reward functions further indicates benefits for more robust reward functions via tighter integration of supervised and preference-based LLM post-training.
comment: Published at NeurIPS 2024
♻ ☆ The Internet of Things in the Era of Generative AI: Vision and Challenges
Advancements in Generative AI hold immense promise to push Internet of Things (IoT) to the next level. In this article, we share our vision on IoT in the era of Generative AI. We discuss some of the most important applications of Generative AI in IoT-related domains. We also identify some of the most critical challenges and discuss current gaps as well as promising opportunities on enabling Generative AI for IoT. We hope this article can inspire new research on IoT in the era of Generative AI.
comment: Featured article of IEEE Internet Computing, 2024, Volume: 28, Issue: 5
♻ ☆ Preventing Representational Rank Collapse in MPNNs by Splitting the Computational Graph
The ability of message-passing neural networks (MPNNs) to fit complex functions over graphs is limited as most graph convolutions amplify the same signal across all feature channels, a phenomenon known as rank collapse, and over-smoothing as a special case. Most approaches to mitigate over-smoothing extend common message-passing schemes, e.g., the graph convolutional network, by utilizing residual connections, gating mechanisms, normalization, or regularization techniques. Our work contrarily proposes to directly tackle the cause of this issue by modifying the message-passing scheme and exchanging different types of messages using multi-relational graphs. We identify a sufficient condition to ensure linearly independent node representations. As one instantion, we show that operating on multiple directed acyclic graphs always satisfies our condition and propose to obtain these by defining a strict partial ordering of the nodes. We conduct comprehensive experiments that confirm the benefits of operating on multi-relational graphs to achieve more informative node representations.
comment: Published at LoG 2024
♻ ☆ Efficient Deep Learning with Decorrelated Backpropagation
The backpropagation algorithm remains the dominant and most successful method for training deep neural networks (DNNs). At the same time, training DNNs at scale comes at a significant computational cost and therefore a high carbon footprint. Converging evidence suggests that input decorrelation may speed up deep learning. However, to date, this has not yet translated into substantial improvements in training efficiency in large-scale DNNs. This is mainly caused by the challenge of enforcing fast and stable network-wide decorrelation. Here, we show for the first time that much more efficient training of very deep neural networks using decorrelated backpropagation is feasible. To achieve this goal we made use of a novel algorithm which induces network-wide input decorrelation using minimal computational overhead. By combining this algorithm with careful optimizations, we obtain a more than two-fold speed-up and higher test accuracy compared to backpropagation when training a 18-layer deep residual network. This demonstrates that decorrelation provides exciting prospects for efficient deep learning at scale.
♻ ☆ Ornstein-Uhlenbeck Adaptation as a Mechanism for Learning in Brains and Machines
Learning is a fundamental property of intelligent systems, observed across biological organisms and engineered systems. While modern intelligent systems typically rely on gradient descent for learning, the need for exact gradients and complex information flow makes its implementation in biological and neuromorphic systems challenging. This has motivated the exploration of alternative learning mechanisms that can operate locally and do not rely on exact gradients. In this work, we introduce a novel approach that leverages noise in the parameters of the system and global reinforcement signals. Using an Ornstein-Uhlenbeck process with adaptive dynamics, our method balances exploration and exploitation during learning, driven by deviations from error predictions, akin to reward prediction error. Operating in continuous time, Orstein-Uhlenbeck adaptation (OUA) is proposed as a general mechanism for learning dynamic, time-evolving environments. We validate our approach across diverse tasks, including supervised learning and reinforcement learning in feedforward and recurrent systems. Additionally, we demonstrate that it can perform meta-learning, adjusting hyper-parameters autonomously. Our results indicate that OUA provides a viable alternative to traditional gradient-based methods, with potential applications in neuromorphic computing. It also hints at a possible mechanism for noise-driven learning in the brain, where stochastic neurotransmitter release may guide synaptic adjustments.
♻ ☆ Faster Convergence on Heterogeneous Federated Edge Learning: An Adaptive Clustered Data Sharing Approach
Federated Edge Learning (FEEL) emerges as a pioneering distributed machine learning paradigm for the 6G Hyper-Connectivity, harnessing data from the Internet of Things (IoT) devices while upholding data privacy. However, current FEEL algorithms struggle with non-independent and non-identically distributed (non-IID) data, leading to elevated communication costs and compromised model accuracy. To address these statistical imbalances within FEEL, we introduce a clustered data sharing framework, mitigating data heterogeneity by selectively sharing partial data from cluster heads to trusted associates through sidelink-aided multicasting. The collective communication pattern is integral to FEEL training, where both cluster formation and the efficiency of communication and computation impact training latency and accuracy simultaneously. To tackle the strictly coupled data sharing and resource optimization, we decompose the overall optimization problem into the clients clustering and effective data sharing subproblems. Specifically, a distribution-based adaptive clustering algorithm (DACA) is devised basing on three deductive cluster forming conditions, which ensures the maximum sharing yield. Meanwhile, we design a stochastic optimization based joint computed frequency and shared data volume optimization (JFVO) algorithm, determining the optimal resource allocation with an uncertain objective function. The experiments show that the proposed framework facilitates FEEL on non-IID datasets with faster convergence rate and higher model accuracy in a limited communication environment.
♻ ☆ Remove that Square Root: A New Efficient Scale-Invariant Version of AdaGrad
Adaptive methods are extremely popular in machine learning as they make learning rate tuning less expensive. This paper introduces a novel optimization algorithm named KATE, which presents a scale-invariant adaptation of the well-known AdaGrad algorithm. We prove the scale-invariance of KATE for the case of Generalized Linear Models. Moreover, for general smooth non-convex problems, we establish a convergence rate of $O \left(\frac{\log T}{\sqrt{T}} \right)$ for KATE, matching the best-known ones for AdaGrad and Adam. We also compare KATE to other state-of-the-art adaptive algorithms Adam and AdaGrad in numerical experiments with different problems, including complex machine learning tasks like image classification and text classification on real data. The results indicate that KATE consistently outperforms AdaGrad and matches/surpasses the performance of Adam in all considered scenarios.
comment: 31 pages, 12 figures
♻ ☆ Speech Robust Bench: A Robustness Benchmark For Speech Recognition NeurIPS
As Automatic Speech Recognition (ASR) models become ever more pervasive, it is important to ensure that they make reliable predictions under corruptions present in the physical and digital world. We propose Speech Robust Bench (SRB), a comprehensive benchmark for evaluating the robustness of ASR models to diverse corruptions. SRB is composed of 114 input perturbations which simulate an heterogeneous range of corruptions that ASR models may encounter when deployed in the wild. We use SRB to evaluate the robustness of several state-of-the-art ASR models and observe that model size and certain modeling choices such as the use of discrete representations, or self-training appear to be conducive to robustness. We extend this analysis to measure the robustness of ASR models on data from various demographic subgroups, namely English and Spanish speakers, and males and females. Our results revealed noticeable disparities in the model's robustness across subgroups. We believe that SRB will significantly facilitate future research towards robust ASR models, by making it easier to conduct comprehensive and comparable robustness evaluations.
comment: submitted to NeurIPS datasets and benchmark track 2025
♻ ☆ LOLA -- An Open-Source Massively Multilingual Large Language Model
This paper presents LOLA, a massively multilingual large language model trained on more than 160 languages using a sparse Mixture-of-Experts Transformer architecture. Our architectural and implementation choices address the challenge of harnessing linguistic diversity while maintaining efficiency and avoiding the common pitfalls of multilinguality. Our analysis of the evaluation results shows competitive performance in natural language generation and understanding tasks. Additionally, we demonstrate how the learned expert-routing mechanism exploits implicit phylogenetic linguistic patterns to potentially alleviate the curse of multilinguality. We provide an in-depth look at the training process, an analysis of the datasets, and a balanced exploration of the model's strengths and limitations. As an open-source model, LOLA promotes reproducibility and serves as a robust foundation for future research. Our findings enable the development of compute-efficient multilingual models with strong, scalable performance across languages.
♻ ☆ Fast Polypharmacy Side Effect Prediction Using Tensor Factorisation
Motivation: Adverse reactions from drug combinations are increasingly common, making their accurate prediction a crucial challenge in modern medicine. Laboratory-based identification of these reactions is insufficient due to the combinatorial nature of the problem. While many computational approaches have been proposed, tensor factorisation models have shown mixed results, necessitating a thorough investigation of their capabilities when properly optimized. Results: We demonstrate that tensor factorisation models can achieve state-of-the-art performance on polypharmacy side effect prediction, with our best model (SimplE) achieving median scores of 0.978 AUROC, 0.971 AUPRC, and 1.000 AP@50 across 963 side effects. Notably, this model reaches 98.3\% of its maximum performance after just two epochs of training (approximately 4 minutes), making it substantially faster than existing approaches while maintaining comparable accuracy. We also find that incorporating monopharmacy data as self-looping edges in the graph performs marginally better than using it to initialize embeddings. Availability and Implementation: All code used in the experiments is available in our GitHub repository (https://doi.org/10.5281/zenodo.10684402). The implementation was carried out using Python 3.8.12 with PyTorch 1.7.1, accelerated with CUDA 11.4 on NVIDIA GeForce RTX 2080 Ti GPUs. Contact: oliver.lloyd@bristol.ac.uk Supplementary information: Supplementary data, including precision-recall curves and F1 curves for the best performing model, are available at Bioinformatics online.
♻ ☆ Group & Reweight: A Novel Cost-Sensitive Approach to Mitigating Class Imbalance in Network Traffic Classification
Internet services have led to the eruption of network traffic, and machine learning on these Internet data has become an indispensable tool, especially when the application is risk-sensitive. This paper focuses on network traffic classification in the presence of class imbalance, which fundamentally and ubiquitously exists in Internet data analysis. This existence of class imbalance mostly drifts the optimal decision boundary and results in a less optimal solution. This brings severe safety concerns in the network traffic field when pattern recognition is challenging with numerous minority malicious classes. To alleviate these effects, we design a \textit{group \& reweight} strategy for alleviating the class imbalance. Inspired by the group distributionally optimization framework, our approach heuristically clusters classes into groups, iteratively updates the non-parametric weights for separate classes and optimizes the learning model by minimizing reweighted losses. We theoretically interpret the optimization process from a Stackelberg game and perform extensive experiments on typical benchmarks. Results show that our approach can not only suppress the negative effect of class imbalance but also improve the comprehensive performance in prediction.
♻ ☆ AGD: an Auto-switchable Optimizer using Stepwise Gradient Difference for Preconditioning Matrix NeurIPS '23
Adaptive optimizers, such as Adam, have achieved remarkable success in deep learning. A key component of these optimizers is the so-called preconditioning matrix, providing enhanced gradient information and regulating the step size of each gradient direction. In this paper, we propose a novel approach to designing the preconditioning matrix by utilizing the gradient difference between two successive steps as the diagonal elements. These diagonal elements are closely related to the Hessian and can be perceived as an approximation of the inner product between the Hessian row vectors and difference of the adjacent parameter vectors. Additionally, we introduce an auto-switching function that enables the preconditioning matrix to switch dynamically between Stochastic Gradient Descent (SGD) and the adaptive optimizer. Based on these two techniques, we develop a new optimizer named AGD that enhances the generalization performance. We evaluate AGD on public datasets of Natural Language Processing (NLP), Computer Vision (CV), and Recommendation Systems (RecSys). Our experimental results demonstrate that AGD outperforms the state-of-the-art (SOTA) optimizers, achieving highly competitive or significantly better predictive performance. Furthermore, we analyze how AGD is able to switch automatically between SGD and the adaptive optimizer and its actual effects on various scenarios. The code is available at https://github.com/intelligent-machine-learning/atorch/tree/main/atorch/optimizers.
comment: 21 pages. Accepted as a conference paper at NeurIPS '23
♻ ☆ Improved Sample Complexity Bounds for Diffusion Model Training
Diffusion models have become the most popular approach to deep generative modeling of images, largely due to their empirical performance and reliability. From a theoretical standpoint, a number of recent works~\cite{chen2022,chen2022improved,benton2023linear} have studied the iteration complexity of sampling, assuming access to an accurate diffusion model. In this work, we focus on understanding the \emph{sample complexity} of training such a model; how many samples are needed to learn an accurate diffusion model using a sufficiently expressive neural network? Prior work~\cite{BMR20} showed bounds polynomial in the dimension, desired Total Variation error, and Wasserstein error. We show an \emph{exponential improvement} in the dependence on Wasserstein error and depth, along with improved dependencies on other relevant parameters.
comment: Bugfix
♻ ☆ PPFlow: Target-aware Peptide Design with Torsional Flow Matching
Therapeutic peptides have proven to have great pharmaceutical value and potential in recent decades. However, methods of AI-assisted peptide drug discovery are not fully explored. To fill the gap, we propose a target-aware peptide design method called \textsc{PPFlow}, based on conditional flow matching on torus manifolds, to model the internal geometries of torsion angles for the peptide structure design. Besides, we establish a protein-peptide binding dataset named PPBench2024 to fill the void of massive data for the task of structure-based peptide drug design and to allow the training of deep learning methods. Extensive experiments show that PPFlow reaches state-of-the-art performance in tasks of peptide drug generation and optimization in comparison with baseline models, and can be generalized to other tasks including docking and side-chain packing.
comment: 18 pages
♻ ☆ Multi-Agent Hybrid SAC for Joint SS-DSA in CRNs
Opportunistic spectrum access has the potential to increase the efficiency of spectrum utilization in cognitive radio networks (CRNs). In CRNs, both spectrum sensing and resource allocation (SSRA) are critical to maximizing system throughput while minimizing collisions of secondary users with the primary network. However, many works in dynamic spectrum access do not consider the impact of imperfect sensing information such as mis-detected channels, which the additional information available in joint SSRA can help remediate. In this work, we examine joint SSRA as an optimization which seeks to maximize a CRN's net communication rate subject to constraints on channel sensing, channel access, and transmit power. Given the non-trivial nature of the problem, we leverage multi-agent reinforcement learning to enable a network of secondary users to dynamically access unoccupied spectrum via only local test statistics, formulated under the energy detection paradigm of spectrum sensing. In doing so, we develop a novel multi-agent implementation of hybrid soft actor critic, MHSAC, based on the QMIX mixing scheme. Through experiments, we find that our SSRA algorithm, HySSRA, is successful in maximizing the CRN's utilization of spectrum resources while also limiting its interference with the primary network, and outperforms the current state-of-the-art by a wide margin. We also explore the impact of wireless variations such as coherence time on the efficacy of the system.
comment: Upon further exploration, model is not converging as expected under current formulation. We are working to update the inputs and objective so that it performs in an expected manner
♻ ☆ Benchmarking the Attribution Quality of Vision Models NeurIPS 2024
Attribution maps are one of the most established tools to explain the functioning of computer vision models. They assign importance scores to input features, indicating how relevant each feature is for the prediction of a deep neural network. While much research has gone into proposing new attribution methods, their proper evaluation remains a difficult challenge. In this work, we propose a novel evaluation protocol that overcomes two fundamental limitations of the widely used incremental-deletion protocol, i.e., the out-of-domain issue and lacking inter-model comparisons. This allows us to evaluate 23 attribution methods and how different design choices of popular vision backbones affect their attribution quality. We find that intrinsically explainable models outperform standard models and that raw attribution values exhibit a higher attribution quality than what is known from previous work. Further, we show consistent changes in the attribution quality when varying the network design, indicating that some standard design choices promote attribution quality.
comment: NeurIPS 2024 Datasets and Benchmarks Track, project page and code: https://github.com/visinf/idsds
♻ ☆ Dynamical similarity analysis uniquely captures how computations develop in RNNs
Methods for analyzing representations in neural systems have become a popular tool in both neuroscience and mechanistic interpretability. Having measures to compare how similar activations of neurons are across conditions, architectures, and species, gives us a scalable way of learning how information is transformed within different neural networks. In contrast to this trend, recent investigations have revealed how some metrics can respond to spurious signals and hence give misleading results. To identify the most reliable metric and understand how measures could be improved, it is going to be important to identify specific test cases which can serve as benchmarks. Here we propose that the phenomena of compositional learning in recurrent neural networks (RNNs) allows us to build a test case for dynamical representation alignment metrics. By implementing this case, we show it enables us to test whether metrics can identify representations which gradually develop throughout learning and probe whether representations identified by metrics are relevant to computations executed by networks. By building both an attractor- and RNN-based test case, we show that the new Dynamical Similarity Analysis (DSA) is more noise robust and identifies behaviorally relevant representations more reliably than prior metrics (Procrustes, CKA). We also show how test cases can be used beyond evaluating metrics to study new architectures. Specifically, results from applying DSA to modern (Mamba) state space models, suggest that, in contrast to RNNs, these models may not exhibit changes to their recurrent dynamics due to their expressiveness. Overall, by developing test cases, we show DSA's exceptional ability to detect compositional dynamical motifs, thereby enhancing our understanding of how computations unfold in RNNs.
comment: 19 pages, 11 figures
♻ ☆ AGBD: A Global-scale Biomass Dataset
Accurate estimates of Above Ground Biomass (AGB) are essential in addressing two of humanity's biggest challenges, climate change and biodiversity loss. Existing datasets for AGB estimation from satellite imagery are limited. Either they focus on specific, local regions at high resolution, or they offer global coverage at low resolution. There is a need for a machine learning-ready, globally representative, high-resolution benchmark. Our findings indicate significant variability in biomass estimates across different vegetation types, emphasizing the necessity for a dataset that accurately captures global diversity. To address these gaps, we introduce a comprehensive new dataset that is globally distributed, covers a range of vegetation types, and spans several years. This dataset combines AGB reference data from the GEDI mission with data from Sentinel-2 and PALSAR-2 imagery. Additionally, it includes pre-processed high-level features such as a dense canopy height map, an elevation map, and a land-cover classification map. We also produce a dense, high-resolution (10m) map of AGB predictions for the entire area covered by the dataset. Rigorously tested, our dataset is accompanied by several benchmark models and is publicly available. It can be easily accessed using a single line of code, offering a solid basis for efforts towards global AGB estimation. The GitHub repository github.com/ghjuliasialelli/AGBD serves as a one-stop shop for all code and data.
♻ ☆ Guiding a Diffusion Model with a Bad Version of Itself NeurIPS 2024
The primary axes of interest in image-generating diffusion models are image quality, the amount of variation in the results, and how well the results align with a given condition, e.g., a class label or a text prompt. The popular classifier-free guidance approach uses an unconditional model to guide a conditional model, leading to simultaneously better prompt alignment and higher-quality images at the cost of reduced variation. These effects seem inherently entangled, and thus hard to control. We make the surprising observation that it is possible to obtain disentangled control over image quality without compromising the amount of variation by guiding generation using a smaller, less-trained version of the model itself rather than an unconditional model. This leads to significant improvements in ImageNet generation, setting record FIDs of 1.01 for 64x64 and 1.25 for 512x512, using publicly available networks. Furthermore, the method is also applicable to unconditional diffusion models, drastically improving their quality.
comment: NeurIPS 2024
♻ ☆ Deterministic Trajectory Optimization through Probabilistic Optimal Control
In this article, we discuss two algorithms tailored to discrete-time deterministic finite-horizon nonlinear optimal control problems or so-called deterministic trajectory optimization problems. Both algorithms can be derived from an emerging theoretical paradigm that we refer to as probabilistic optimal control. The paradigm reformulates stochastic optimal control as an equivalent probabilistic inference problem and can be viewed as a generalisation of the former. The merit of this perspective is that it allows to address the problem using the Expectation-Maximization algorithm. It is shown that the application of this algorithm results in a fixed point iteration of probabilistic policies that converge to the deterministic optimal policy. Two strategies for policy evaluation are discussed, using state-of-the-art uncertainty quantification methods resulting into two distinct algorithms. The algorithms are structurally closest related to the differential dynamic programming algorithm and related methods that use sigma-point methods to avoid direct gradient evaluations. The main advantage of the algorithms is an improved balance between exploration and exploitation over the iterations, leading to improved numerical stability and accelerated convergence. These properties are demonstrated on different nonlinear systems.
♻ ☆ Identifying percolation phase transitions with unsupervised learning based on largest clusters
The application of machine learning in the study of phase transitions has achieved remarkable success in both equilibrium and non-equilibrium systems. It is widely recognized that unsupervised learning can retrieve phase transition information through hidden variables. However, using unsupervised methods to identify the critical point of percolation models has remained an intriguing challenge. This paper suggests that, by inputting the largest cluster rather than the original configuration into the learning model, unsupervised learning can indeed predict the critical point of the percolation model. Furthermore, we observe that when the largest cluster configuration is randomly shuffled-altering the positions of occupied sites or bonds-there is no significant difference in the output compared to learning the largest cluster configuration directly. This finding suggests a more general principle: unsupervised learning primarily captures particle density, or more specifically, occupied site density. However, shuffling does impact the formation of the largest cluster, which is directly related to phase transitions. As randomness increases, we observe that the correlation length tends to decrease, providing direct evidence of this relationship. We also propose a method called Fake Finite Size Scaling (FFSS) to calculate the critical value, which improves the accuracy of fitting to a great extent.
comment: 15 pages,40 figures,39 references
♻ ☆ SPEAR:Exact Gradient Inversion of Batches in Federated Learning
Federated learning is a framework for collaborative machine learning where clients only share gradient updates and not their private data with a server. However, it was recently shown that gradient inversion attacks can reconstruct this data from the shared gradients. In the important honest-but-curious setting, existing attacks enable exact reconstruction only for batch size of $b=1$, with larger batches permitting only approximate reconstruction. In this work, we propose SPEAR, the first algorithm reconstructing whole batches with $b >1$ exactly. SPEAR combines insights into the explicit low-rank structure of gradients with a sampling-based algorithm. Crucially, we leverage ReLU-induced gradient sparsity to precisely filter out large numbers of incorrect samples, making a final reconstruction step tractable. We provide an efficient GPU implementation for fully connected networks and show that it recovers high-dimensional ImageNet inputs in batches of up to $b \lesssim 25$ exactly while scaling to large networks. Finally, we show theoretically that much larger batches can be reconstructed with high probability given exponential time.
♻ ☆ Dataset Distribution Impacts Model Fairness: Single vs. Multi-Task Learning
The influence of bias in datasets on the fairness of model predictions is a topic of ongoing research in various fields. We evaluate the performance of skin lesion classification using ResNet-based CNNs, focusing on patient sex variations in training data and three different learning strategies. We present a linear programming method for generating datasets with varying patient sex and class labels, taking into account the correlations between these variables. We evaluated the model performance using three different learning strategies: a single-task model, a reinforcing multi-task model, and an adversarial learning scheme. Our observations include: 1) sex-specific training data yields better results, 2) single-task models exhibit sex bias, 3) the reinforcement approach does not remove sex bias, 4) the adversarial model eliminates sex bias in cases involving only female patients, and 5) datasets that include male patients enhance model performance for the male subgroup, even when female patients are the majority. To generalise these findings, in future research, we will examine more demographic attributes, like age, and other possibly confounding factors, such as skin colour and artefacts in the skin lesions. We make all data and models available on GitHub.
comment: Published in the FAIMI EPIMI 2024 Workshop
♻ ☆ Generator Matching: Generative modeling with arbitrary Markov processes
We introduce generator matching, a modality-agnostic framework for generative modeling using arbitrary Markov processes. Generators characterize the infinitesimal evolution of a Markov process, which we leverage for generative modeling in a similar vein to flow matching: we construct conditional generators which generate single data points, then learn to approximate the marginal generator which generates the full data distribution. We show that generator matching unifies various generative modeling methods, including diffusion models, flow matching and discrete diffusion models. Furthermore, it provides the foundation to expand the design space to new and unexplored Markov processes such as jump processes. Finally, generator matching enables the construction of superpositions of Markov generative processes and enables the construction of multimodal models in a rigorous manner. We empirically validate our method on protein and image structure generation, showing that superposition with a jump process improves image generation.
♻ ☆ Task Groupings Regularization: Data-Free Meta-Learning with Heterogeneous Pre-trained Models
Data-Free Meta-Learning (DFML) aims to derive knowledge from a collection of pre-trained models without accessing their original data, enabling the rapid adaptation to new unseen tasks. Current methods often overlook the heterogeneity among pre-trained models, which leads to performance degradation due to task conflicts. In this paper, we empirically and theoretically identify and analyze the model heterogeneity in DFML. We find that model heterogeneity introduces a heterogeneity-homogeneity trade-off, where homogeneous models reduce task conflicts but also increase the overfitting risk. Balancing this trade-off is crucial for learning shared representations across tasks. Based on our findings, we propose Task Groupings Regularization that benefits from model heterogeneity by grouping and aligning conflicting tasks. Specifically, we embed pre-trained models into a task space to compute dissimilarity, and group heterogeneous models together based on this measure. Then, we introduce implicit gradient regularization within each group to mitigate potential conflicts. By encouraging a gradient direction suitable for all tasks, the meta-model captures shared representations that generalize across tasks. Comprehensive experiments showcase the superiority of our approach in multiple benchmarks, effectively tackling the model heterogeneity in challenging multi-domain and multi-architecture scenarios.
♻ ☆ psifx -- Psychological and Social Interactions Feature Extraction Package
psifx is a plug-and-play multi-modal feature extraction toolkit, aiming to facilitate and democratize the use of state-of-the-art machine learning techniques for human sciences research. It is motivated by a need (a) to automate and standardize data annotation processes, otherwise involving expensive, lengthy, and inconsistent human labor, such as the transcription or coding of behavior changes from audio and video sources; (b) to develop and distribute open-source community-driven psychology research software; and (c) to enable large-scale access and ease of use to non-expert users. The framework contains an array of tools for tasks, such as speaker diarization, closed-caption transcription and translation from audio, as well as body, hand, and facial pose estimation and gaze tracking from video. The package has been designed with a modular and task-oriented approach, enabling the community to add or update new tools easily. We strongly hope that this package will provide psychologists a simple and practical solution for efficiently a range of audio, linguistic, and visual features from audio and video, thereby creating new opportunities for in-depth study of real-time behavioral phenomena.
♻ ☆ Learning to Simulate: Generative Metamodeling via Quantile Regression
Stochastic simulation models effectively capture complex system dynamics but are often too slow for real-time decision-making. Traditional metamodeling techniques learn relationships between simulator inputs and a single output summary statistic, such as the mean or median. These techniques enable real-time predictions without additional simulations. However, they require prior selection of one appropriate output summary statistic, limiting their flexibility in practical applications. We propose a new concept: generative metamodeling. It aims to construct a "fast simulator of the simulator," generating random outputs significantly faster than the original simulator while preserving approximately equal conditional distributions. Generative metamodels enable rapid generation of numerous random outputs upon input specification, facilitating immediate computation of any summary statistic for real-time decision-making. We introduce a new algorithm, quantile-regression-based generative metamodeling (QRGMM), and establish its distributional convergence and convergence rate. Extensive numerical experiments demonstrate QRGMM's efficacy compared to other state-of-the-art generative algorithms in practical real-time decision-making scenarios.
comment: Main body: 42 pages, 10 figures; supplemental material: 22 pages
♻ ☆ FREE: Faster and Better Data-Free Meta-Learning
Data-Free Meta-Learning (DFML) aims to extract knowledge from a collection of pre-trained models without requiring the original data, presenting practical benefits in contexts constrained by data privacy concerns. Current DFML methods primarily focus on the data recovery from these pre-trained models. However, they suffer from slow recovery speed and overlook gaps inherent in heterogeneous pre-trained models. In response to these challenges, we introduce the Faster and Better Data-Free Meta-Learning (FREE) framework, which contains: (i) a meta-generator for rapidly recovering training tasks from pre-trained models; and (ii) a meta-learner for generalizing to new unseen tasks. Specifically, within the module Faster Inversion via Meta-Generator, each pre-trained model is perceived as a distinct task. The meta-generator can rapidly adapt to a specific task in just five steps, significantly accelerating the data recovery. Furthermore, we propose Better Generalization via Meta-Learner and introduce an implicit gradient alignment algorithm to optimize the meta-learner. This is achieved as aligned gradient directions alleviate potential conflicts among tasks from heterogeneous pre-trained models. Empirical experiments on multiple benchmarks affirm the superiority of our approach, marking a notable speed-up (20$\times$) and performance enhancement (1.42%$\sim$4.78%) in comparison to the state-of-the-art.
♻ ☆ Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Models NeurIPS 2024
To build intelligent machine learning systems, there are two broad approaches. One approach is to build inherently interpretable models, as endeavored by the growing field of causal representation learning. The other approach is to build highly-performant foundation models and then invest efforts into understanding how they work. In this work, we relate these two approaches and study how to learn human-interpretable concepts from data. Weaving together ideas from both fields, we formally define a notion of concepts and show that they can be provably recovered from diverse data. Experiments on synthetic data and large language models show the utility of our unified approach.
comment: To appear in NeurIPS 2024 under the modified title 'From Causal to Concept-Based Representation Learning'
♻ ☆ TLDR: Unsupervised Goal-Conditioned RL via Temporal Distance-Aware Representations
Unsupervised goal-conditioned reinforcement learning (GCRL) is a promising paradigm for developing diverse robotic skills without external supervision. However, existing unsupervised GCRL methods often struggle to cover a wide range of states in complex environments due to their limited exploration and sparse or noisy rewards for GCRL. To overcome these challenges, we propose a novel unsupervised GCRL method that leverages TemporaL Distance-aware Representations (TLDR). Based on temporal distance, TLDR selects faraway goals to initiate exploration and computes intrinsic exploration rewards and goal-reaching rewards. Specifically, our exploration policy seeks states with large temporal distances (i.e. covering a large state space), while the goal-conditioned policy learns to minimize the temporal distance to the goal (i.e. reaching the goal). Our results in six simulated locomotion environments demonstrate that TLDR significantly outperforms prior unsupervised GCRL methods in achieving a wide range of states.
comment: CoRL 2024
♻ ☆ FairML: A Julia Package for Fair Classification
In this paper, we propose FairML.jl, a Julia package providing a framework for fair classification in machine learning. In this framework, the fair learning process is divided into three stages. Each stage aims to reduce unfairness, such as disparate impact and disparate mistreatment, in the final prediction. For the preprocessing stage, we present a resampling method that addresses unfairness coming from data imbalances. The in-processing phase consist of a classification method. This can be either one coming from the MLJ.jl package, or a user defined one. For this phase, we incorporate fair ML methods that can handle unfairness to a certain degree through their optimization process. In the post-processing, we discuss the choice of the cut-off value for fair prediction. With simulations, we show the performance of the single phases and their combinations.
comment: 25 pages, 8 figures
♻ ☆ Fair Generalized Linear Mixed Models
When using machine learning for automated prediction, it is important to account for fairness in the prediction. Fairness in machine learning aims to ensure that biases in the data and model inaccuracies do not lead to discriminatory decisions. E.g., predictions from fair machine learning models should not discriminate against sensitive variables such as sexual orientation and ethnicity. The training data often in obtained from social surveys. In social surveys, oftentimes the data collection process is a strata sampling, e.g. due to cost restrictions. In strata samples, the assumption of independence between the observation is not fulfilled. Hence, if the machine learning models do not account for the strata correlations, the results may be biased. Especially high is the bias in cases where the strata assignment is correlated to the variable of interest. We present in this paper an algorithm that can handle both problems simultaneously, and we demonstrate the impact of stratified sampling on the quality of fair machine learning predictions in a reproducible simulation study.
comment: 25 pages, 12 figures. arXiv admin note: text overlap with arXiv:2405.06433
♻ ☆ An Efficient Loop and Clique Coarsening Algorithm for Graph Classification
Graph Transformers (GTs) have made remarkable achievements in graph-level tasks. However, most existing works regard graph structures as a form of guidance or bias for enhancing node representations, which focuses on node-central perspectives and lacks explicit representations of edges and structures. One natural question arises as to whether we can leverage a hypernode to represent some structures. Through experimental analysis, we explore the feasibility of this assumption. Based on our findings, we propose an efficient Loop and Clique Coarsening algorithm with linear complexity for Graph Classification (LCC4GC) on GT architecture. Specifically, we build three unique views, original, coarsening, and conversion, to learn a thorough structural representation. We compress loops and cliques via hierarchical heuristic graph coarsening and restrict them with well-designed constraints, which builds the coarsening view to learn high-level interactions between structures. We also introduce line graphs for edge embeddings and switch to edge-central perspective to alleviate the impact of coarsening reduction. Experiments on eight real-world datasets demonstrate the improvements of LCC4GC over 31 baselines from various architectures.
♻ ☆ Closing the Gap: Achieving Global Convergence (Last Iterate) of Actor-Critic under Markovian Sampling with Neural Network Parametrization ICML 2024
The current state-of-the-art theoretical analysis of Actor-Critic (AC) algorithms significantly lags in addressing the practical aspects of AC implementations. This crucial gap needs bridging to bring the analysis in line with practical implementations of AC. To address this, we advocate for considering the MMCLG criteria: \textbf{M}ulti-layer neural network parametrization for actor/critic, \textbf{M}arkovian sampling, \textbf{C}ontinuous state-action spaces, the performance of the \textbf{L}ast iterate, and \textbf{G}lobal optimality. These aspects are practically significant and have been largely overlooked in existing theoretical analyses of AC algorithms. In this work, we address these gaps by providing the first comprehensive theoretical analysis of AC algorithms that encompasses all five crucial practical aspects (covers MMCLG criteria). We establish global convergence sample complexity bounds of $\tilde{\mathcal{O}}\left({\epsilon^{-3}}\right)$. We achieve this result through our novel use of the weak gradient domination property of MDP's and our unique analysis of the error in critic estimation.
comment: Accepted at ICML 2024. This is a revised version of arXiv:2306.10486, where we have gone from finite action space to continuous action space, from average iterate convergence to last iterate convergence and from $\epsilon^{-4}$ to $\epsilon^{-3}$ sample complexity. This version fixes the related work result of (Xu et al., 2020a), based on their result update on arXiv
♻ ☆ Auditing Fairness under Unobserved Confounding AISTATS 2024
Many definitions of fairness or inequity involve unobservable causal quantities that cannot be directly estimated without strong assumptions. For instance, it is particularly difficult to estimate notions of fairness that rely on hard-to-measure concepts such as risk (e.g., quantifying whether patients at the same risk level have equal probability of treatment, regardless of group membership). Such measurements of risk can be accurately obtained when no unobserved confounders have jointly influenced past decisions and outcomes. However, in the real world, this assumption rarely holds. In this paper, we show that, surprisingly, one can still compute meaningful bounds on treatment rates for high-risk individuals (i.e., conditional on their true, \textit{unobserved} negative outcome), even when entirely eliminating or relaxing the assumption that we observe all relevant risk factors used by decision makers. We use the fact that in many real-world settings (e.g., the release of a new treatment) we have data from prior to any allocation to derive unbiased estimates of risk. This result enables us to audit unfair outcomes of existing decision-making systems in a principled manner. We demonstrate the effectiveness of our framework with a real-world study of Paxlovid allocation, provably identifying that observed racial inequity cannot be explained by unobserved confounders of the same strength as important observed covariates.
comment: AISTATS 2024
♻ ☆ Space to Policy: Scalable Brick Kiln Detection and Automatic Compliance Monitoring with Geospatial Data
Air pollution kills 7 million people annually. The brick kiln sector significantly contributes to economic development but also accounts for 8-14\% of air pollution in India. Policymakers have implemented compliance measures to regulate brick kilns. Emission inventories are critical for air quality modeling and source apportionment studies. However, the largely unorganized nature of the brick kiln sector necessitates labor-intensive survey efforts for monitoring. Recent efforts by air quality researchers have relied on manual annotation of brick kilns using satellite imagery to build emission inventories, but this approach lacks scalability. Machine-learning-based object detection methods have shown promise for detecting brick kilns; however, previous studies often rely on costly high-resolution imagery and fail to integrate with governmental policies. In this work, we developed a scalable machine-learning pipeline that detected and classified 30638 brick kilns across five states in the Indo-Gangetic Plain using free, moderate-resolution satellite imagery from Planet Labs. Our detections have a high correlation with on-ground surveys. We performed automated compliance analysis based on government policies. In the Delhi airshed, stricter policy enforcement has led to the adoption of efficient brick kiln technologies. This study highlights the need for inclusive policies that balance environmental sustainability with the livelihoods of workers.
♻ ☆ Grounding Multimodal Large Language Models in Actions
Multimodal Large Language Models (MLLMs) have demonstrated a wide range of capabilities across many domains, including Embodied AI. In this work, we study how to best ground a MLLM into different embodiments and their associated action spaces, with the goal of leveraging the multimodal world knowledge of the MLLM. We first generalize a number of methods through a unified architecture and the lens of action space adaptors. For continuous actions, we show that a learned tokenization allows for sufficient modeling precision, yielding the best performance on downstream tasks. For discrete actions, we demonstrate that semantically aligning these actions with the native output token space of the MLLM leads to the strongest performance. We arrive at these lessons via a thorough study of seven action space adapters on five different environments, encompassing over 114 embodied tasks.
♻ ☆ Forte : Finding Outliers with Representation Typicality Estimation
Generative models can now produce photorealistic synthetic data which is virtually indistinguishable from the real data used to train it. This is a significant evolution over previous models which could produce reasonable facsimiles of the training data, but ones which could be visually distinguished from the training data by human evaluation. Recent work on OOD detection has raised doubts that generative model likelihoods are optimal OOD detectors due to issues involving likelihood misestimation, entropy in the generative process, and typicality. We speculate that generative OOD detectors also failed because their models focused on the pixels rather than the semantic content of the data, leading to failures in near-OOD cases where the pixels may be similar but the information content is significantly different. We hypothesize that estimating typical sets using self-supervised learners leads to better OOD detectors. We introduce a novel approach that leverages representation learning, and informative summary statistics based on manifold estimation, to address all of the aforementioned issues. Our method outperforms other unsupervised approaches and achieves state-of-the art performance on well-established challenging benchmarks, and new synthetic data detection tasks.
♻ ☆ Rethinking Reward Model Evaluation: Are We Barking up the Wrong Tree?
Reward Models (RMs) are crucial for aligning language models with human preferences. Currently, the evaluation of RMs depends on measuring accuracy against a validation set of manually annotated preference data. Although this method is straightforward and widely adopted, the relationship between RM accuracy and downstream policy performance remains under-explored. In this work, we conduct experiments in a synthetic setting to investigate how differences in RM measured by accuracy translate into gaps in optimized policy performance. Our findings reveal that while there is a weak positive correlation between accuracy and downstream performance, policies optimized towards RMs with similar accuracy can exhibit quite different performance. Moreover, we discover that the way of measuring accuracy significantly impacts its ability to predict the final policy performance. Through the lens of the Regressional Goodhart effect, we recognize that accuracy, when used for measuring RM quality, can fail to fully capture the potential RM overoptimization. This underscores the inadequacy of relying solely on accuracy to reflect their impact on policy optimization.
Artificial Intelligence 148
☆ [MASK] is All You Need
In generative models, two paradigms have gained attraction in various applications: next-set prediction-based Masked Generative Models and next-noise prediction-based Non-Autoregressive Models, e.g., Diffusion Models. In this work, we propose using discrete-state models to connect them and explore their scalability in the vision domain. First, we conduct a step-by-step analysis in a unified design space across two types of models including timestep-independence, noise schedule, temperature, guidance strength, etc in a scalable manner. Second, we re-cast typical discriminative tasks, e.g., image segmentation, as an unmasking process from [MASK]tokens on a discrete-state model. This enables us to perform various sampling processes, including flexible conditional sampling by only training once to model the joint distribution. All aforementioned explorations lead to our framework named Discrete Interpolants, which enables us to achieve state-of-the-art or competitive performance compared to previous discrete-state based methods in various benchmarks, like ImageNet256, MS COCO, and video dataset FaceForensics. In summary, by leveraging [MASK] in discrete-state models, we can bridge Masked Generative and Non-autoregressive Diffusion models, as well as generative and discriminative tasks.
comment: Technical Report (WIP), Project Page(code, model, dataset): https://compvis.github.io/mask/
☆ P3-PO: Prescriptive Point Priors for Visuo-Spatial Generalization of Robot Policies
Developing generalizable robot policies that can robustly handle varied environmental conditions and object instances remains a fundamental challenge in robot learning. While considerable efforts have focused on collecting large robot datasets and developing policy architectures to learn from such data, naively learning from visual inputs often results in brittle policies that fail to transfer beyond the training data. This work presents Prescriptive Point Priors for Policies or P3-PO, a novel framework that constructs a unique state representation of the environment leveraging recent advances in computer vision and robot learning to achieve improved out-of-distribution generalization for robot manipulation. This representation is obtained through two steps. First, a human annotator prescribes a set of semantically meaningful points on a single demonstration frame. These points are then propagated through the dataset using off-the-shelf vision models. The derived points serve as an input to state-of-the-art policy architectures for policy learning. Our experiments across four real-world tasks demonstrate an overall 43% absolute improvement over prior methods when evaluated in identical settings as training. Further, P3-PO exhibits 58% and 80% gains across tasks for new object instances and more cluttered environments respectively. Videos illustrating the robot's performance are best viewed at point-priors.github.io.
☆ AnyBimanual: Transferring Unimanual Policy for General Bimanual Manipulation
Performing general language-conditioned bimanual manipulation tasks is of great importance for many applications ranging from household service to industrial assembly. However, collecting bimanual manipulation data is expensive due to the high-dimensional action space, which poses challenges for conventional methods to handle general bimanual manipulation tasks. In contrast, unimanual policy has recently demonstrated impressive generalizability across a wide range of tasks because of scaled model parameters and training data, which can provide sharable manipulation knowledge for bimanual systems. To this end, we propose a plug-and-play method named AnyBimanual, which transfers pre-trained unimanual policy to general bimanual manipulation policy with few bimanual demonstrations. Specifically, we first introduce a skill manager to dynamically schedule the skill representations discovered from pre-trained unimanual policy for bimanual manipulation tasks, which linearly combines skill primitives with task-oriented compensation to represent the bimanual manipulation instruction. To mitigate the observation discrepancy between unimanual and bimanual systems, we present a visual aligner to generate soft masks for visual embedding of the workspace, which aims to align visual input of unimanual policy model for each arm with those during pretraining stage. AnyBimanual shows superiority on 12 simulated tasks from RLBench2 with a sizable 12.67% improvement in success rate over previous methods. Experiments on 9 real-world tasks further verify its practicality with an average success rate of 84.62%.
comment: Project page: https://anybimanual.github.io/
☆ Driv3R: Learning Dense 4D Reconstruction for Autonomous Driving
Realtime 4D reconstruction for dynamic scenes remains a crucial challenge for autonomous driving perception. Most existing methods rely on depth estimation through self-supervision or multi-modality sensor fusion. In this paper, we propose Driv3R, a DUSt3R-based framework that directly regresses per-frame point maps from multi-view image sequences. To achieve streaming dense reconstruction, we maintain a memory pool to reason both spatial relationships across sensors and dynamic temporal contexts to enhance multi-view 3D consistency and temporal integration. Furthermore, we employ a 4D flow predictor to identify moving objects within the scene to direct our network focus more on reconstructing these dynamic regions. Finally, we align all per-frame pointmaps consistently to the world coordinate system in an optimization-free manner. We conduct extensive experiments on the large-scale nuScenes dataset to evaluate the effectiveness of our method. Driv3R outperforms previous frameworks in 4D dynamic scene reconstruction, achieving 15x faster inference speed compared to methods requiring global alignment. Code: https://github.com/Barrybarry-Smith/Driv3R.
comment: Code is available at: https://github.com/Barrybarry-Smith/Driv3R
☆ Delve into Visual Contrastive Decoding for Hallucination Mitigation of Large Vision-Language Models
While large vision-language models (LVLMs) have shown impressive capabilities in generating plausible responses correlated with input visual contents, they still suffer from hallucinations, where the generated text inaccurately reflects visual contents. To address this, recent approaches apply contrastive decoding to calibrate the model's response via contrasting output distributions with original and visually distorted samples, demonstrating promising hallucination mitigation in a training-free manner. However, the potential of changing information in visual inputs is not well-explored, so a deeper investigation into the behaviors of visual contrastive decoding is of great interest. In this paper, we first explore various methods for contrastive decoding to change visual contents, including image downsampling and editing. Downsampling images reduces the detailed textual information while editing yields new contents in images, providing new aspects as visual contrastive samples. To further study benefits by using different contrastive samples, we analyze probability-level metrics, including entropy and distribution distance. Interestingly, the effect of these samples in mitigating hallucinations varies a lot across LVLMs and benchmarks. Based on our analysis, we propose a simple yet effective method to combine contrastive samples, offering a practical solution for applying contrastive decoding across various scenarios. Extensive experiments are conducted to validate the proposed fusion method among different benchmarks.
comment: Under review. Project pages: https://github.com/YiLunLee/VCD_Analysis
☆ Visual Lexicon: Rich Image Features in Language Space
We present Visual Lexicon, a novel visual language that encodes rich image information into the text space of vocabulary tokens while retaining intricate visual details that are often challenging to convey in natural language. Unlike traditional methods that prioritize either high-level semantics (e.g., CLIP) or pixel-level reconstruction (e.g., VAE), ViLex simultaneously captures rich semantic content and fine visual details, enabling high-quality image generation and comprehensive visual scene understanding. Through a self-supervised learning pipeline, ViLex generates tokens optimized for reconstructing input images using a frozen text-to-image (T2I) diffusion model, preserving the detailed information necessary for high-fidelity semantic-level reconstruction. As an image embedding in the language space, ViLex tokens leverage the compositionality of natural languages, allowing them to be used independently as "text tokens" or combined with natural language tokens to prompt pretrained T2I models with both visual and textual inputs, mirroring how we interact with vision-language models (VLMs). Experiments demonstrate that ViLex achieves higher fidelity in image reconstruction compared to text embeddings--even with a single ViLex token. Moreover, ViLex successfully performs various DreamBooth tasks in a zero-shot, unsupervised manner without fine-tuning T2I models. Additionally, ViLex serves as a powerful vision encoder, consistently improving vision-language model performance across 15 benchmarks relative to a strong SigLIP baseline.
comment: Tech report. 16 pages, 10 figures
☆ Proactive Agents for Multi-Turn Text-to-Image Generation Under Uncertainty
User prompts for generative AI models are often underspecified, leading to sub-optimal responses. This problem is particularly evident in text-to-image (T2I) generation, where users commonly struggle to articulate their precise intent. This disconnect between the user's vision and the model's interpretation often forces users to painstakingly and repeatedly refine their prompts. To address this, we propose a design for proactive T2I agents equipped with an interface to (1) actively ask clarification questions when uncertain, and (2) present their understanding of user intent as an understandable belief graph that a user can edit. We build simple prototypes for such agents and verify their effectiveness through both human studies and automated evaluation. We observed that at least 90% of human subjects found these agents and their belief graphs helpful for their T2I workflow. Moreover, we develop a scalable automated evaluation approach using two agents, one with a ground truth image and the other tries to ask as few questions as possible to align with the ground truth. On DesignBench, a benchmark we created for artists and designers, the COCO dataset (Lin et al., 2014), and ImageInWords (Garg et al., 2024), we observed that these T2I agents were able to ask informative questions and elicit crucial information to achieve successful alignment with at least 2 times higher VQAScore (Lin et al., 2024) than the standard single-turn T2I generation. Demo: https://github.com/google-deepmind/proactive_t2i_agents.
☆ XRZoo: A Large-Scale and Versatile Dataset of Extended Reality (XR) Applications
The rapid advancement of Extended Reality (XR, encompassing AR, MR, and VR) and spatial computing technologies forms a foundational layer for the emerging Metaverse, enabling innovative applications across healthcare, education, manufacturing, and entertainment. However, research in this area is often limited by the lack of large, representative, and highquality application datasets that can support empirical studies and the development of new approaches benefiting XR software processes. In this paper, we introduce XRZoo, a comprehensive and curated dataset of XR applications designed to bridge this gap. XRZoo contains 12,528 free XR applications, spanning nine app stores, across all XR techniques (i.e., AR, MR, and VR) and use cases, with detailed metadata on key aspects such as application descriptions, application categories, release dates, user review numbers, and hardware specifications, etc. By making XRZoo publicly available, we aim to foster reproducible XR software engineering and security research, enable cross-disciplinary investigations, and also support the development of advanced XR systems by providing examples to developers. Our dataset serves as a valuable resource for researchers and practitioners interested in improving the scalability, usability, and effectiveness of XR applications. XRZoo will be released and actively maintained.
☆ ContRail: A Framework for Realistic Railway Image Synthesis using ControlNet
Deep Learning became an ubiquitous paradigm due to its extraordinary effectiveness and applicability in numerous domains. However, the approach suffers from the high demand of data required to achieve the potential of this type of model. An ever-increasing sub-field of Artificial Intelligence, Image Synthesis, aims to address this limitation through the design of intelligent models capable of creating original and realistic images, endeavour which could drastically reduce the need for real data. The Stable Diffusion generation paradigm recently propelled state-of-the-art approaches to exceed all previous benchmarks. In this work, we propose the ContRail framework based on the novel Stable Diffusion model ControlNet, which we empower through a multi-modal conditioning method. We experiment with the task of synthetic railway image generation, where we improve the performance in rail-specific tasks, such as rail semantic segmentation by enriching the dataset with realistic synthetic images.
comment: 9 pages, 5 figures, 2 tables
☆ Toward Non-Invasive Diagnosis of Bankart Lesions with Deep Learning SP
Bankart lesions, or anterior-inferior glenoid labral tears, are diagnostically challenging on standard MRIs due to their subtle imaging features-often necessitating invasive MRI arthrograms (MRAs). This study develops deep learning (DL) models to detect Bankart lesions on both standard MRIs and MRAs, aiming to improve diagnostic accuracy and reduce reliance on MRAs. We curated a dataset of 586 shoulder MRIs (335 standard, 251 MRAs) from 558 patients who underwent arthroscopy. Ground truth labels were derived from intraoperative findings, the gold standard for Bankart lesion diagnosis. Separate DL models for MRAs and standard MRIs were trained using the Swin Transformer architecture, pre-trained on a public knee MRI dataset. Predictions from sagittal, axial, and coronal views were ensembled to optimize performance. The models were evaluated on a 20% hold-out test set (117 MRIs: 46 MRAs, 71 standard MRIs). Bankart lesions were identified in 31.9% of MRAs and 8.6% of standard MRIs. The models achieved AUCs of 0.87 (86% accuracy, 83% sensitivity, 86% specificity) and 0.90 (85% accuracy, 82% sensitivity, 86% specificity) on standard MRIs and MRAs, respectively. These results match or surpass radiologist performance on our dataset and reported literature metrics. Notably, our model's performance on non-invasive standard MRIs matched or surpassed the radiologists interpreting MRAs. This study demonstrates the feasibility of using DL to address the diagnostic challenges posed by subtle pathologies like Bankart lesions. Our models demonstrate potential to improve diagnostic confidence, reduce reliance on invasive imaging, and enhance accessibility to care.
comment: Accepted for presentation at SPIE Medical Imaging 2025: Computer-Aided Diagnosis. The manuscript is expected to appear in the conference proceedings
☆ Parkinson's Disease Diagnosis Through Deep Learning: A Novel LSTM-Based Approach for Freezing of Gait Detection
Deep learning holds tremendous potential in healthcare for uncovering hidden patterns within extensive clinical datasets, aiding in the diagnosis of various diseases. Parkinson's disease (PD) is a neurodegenerative condition characterized by the deterioration of brain function. In the initial stages of PD, automatic diagnosis poses a challenge due to the similarity in behavior between individuals with PD and those who are healthy. Our objective is to propose an effective model that can aid in the early detection of Parkinson's disease. We employed the VGRF gait signal dataset sourced from Physionet for distinguishing between healthy individuals and those diagnosed with Parkinson's disease. This paper introduces a novel deep learning architecture based on the LSTM network for automatically detecting freezing of gait episodes in Parkinson's disease patients. In contrast to conventional machine learning algorithms, this method eliminates manual feature engineering and proficiently captures prolonged temporal dependencies in gait patterns, thereby improving the diagnosis of Parkinson's disease. The LSTM network resolves the issue of vanishing gradients by employing memory blocks in place of self-connected hidden units, allowing for optimal information assimilation. To prevent overfitting, dropout and L2 regularization techniques have been employed. Additionally, the stochastic gradient-based optimizer Adam is used for the optimization process. The results indicate that our proposed approach surpasses current state-of-the-art models in FOG episode detection, achieving an accuracy of 97.71%, sensitivity of 99%, precision of 98%, and specificity of 96%. This demonstrates its potential as a superior classification method for Parkinson's disease detection.
☆ Source Separation & Automatic Transcription for Music
Source separation is the process of isolating individual sounds in an auditory mixture of multiple sounds [1], and has a variety of applications ranging from speech enhancement and lyric transcription [2] to digital audio production for music. Furthermore, Automatic Music Transcription (AMT) is the process of converting raw music audio into sheet music that musicians can read [3]. Historically, these tasks have faced challenges such as significant audio noise, long training times, and lack of free-use data due to copyright restrictions. However, recent developments in deep learning have brought new promising approaches to building low-distortion stems and generating sheet music from audio signals [4]. Using spectrogram masking, deep neural networks, and the MuseScore API, we attempt to create an end-to-end pipeline that allows for an initial music audio mixture (e.g...wav file) to be separated into instrument stems, converted into MIDI files, and transcribed into sheet music for each component instrument.
☆ Digital Transformation in the Water Distribution System based on the Digital Twins Concept
Digital Twins have emerged as a disruptive technology with great potential; they can enhance WDS by offering real-time monitoring, predictive maintenance, and optimization capabilities. This paper describes the development of a state-of-the-art DT platform for WDS, introducing advanced technologies such as the Internet of Things, Artificial Intelligence, and Machine Learning models. This paper provides insight into the architecture of the proposed platform-CAUCCES-that, informed by both historical and meteorological data, effectively deploys AI/ML models like LSTM networks, Prophet, LightGBM, and XGBoost in trying to predict water consumption patterns. Furthermore, we delve into how optimization in the maintenance of WDS can be achieved by formulating a Constraint Programming problem for scheduling, hence minimizing the operational cost efficiently with reduced environmental impacts. It also focuses on cybersecurity and protection to ensure the integrity and reliability of the DT platform. In this view, the system will contribute to improvements in decision-making capabilities, operational efficiency, and system reliability, with reassurance being drawn from the important role it can play toward sustainable management of water resources.
comment: 78 pages, 18 figures
☆ OmniEvalKit: A Modular, Lightweight Toolbox for Evaluating Large Language Model and its Omni-Extensions
The rapid advancements in Large Language Models (LLMs) have significantly expanded their applications, ranging from multilingual support to domain-specific tasks and multimodal integration. In this paper, we present OmniEvalKit, a novel benchmarking toolbox designed to evaluate LLMs and their omni-extensions across multilingual, multidomain, and multimodal capabilities. Unlike existing benchmarks that often focus on a single aspect, OmniEvalKit provides a modular, lightweight, and automated evaluation system. It is structured with a modular architecture comprising a Static Builder and Dynamic Data Flow, promoting the seamless integration of new models and datasets. OmniEvalKit supports over 100 LLMs and 50 evaluation datasets, covering comprehensive evaluations across thousands of model-dataset combinations. OmniEvalKit is dedicated to creating an ultra-lightweight and fast-deployable evaluation framework, making downstream applications more convenient and versatile for the AI community.
☆ Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone
Recent advances in learning decision-making policies can largely be attributed to training expressive policy models, largely via imitation learning. While imitation learning discards non-expert data, reinforcement learning (RL) can still learn from suboptimal data. However, instantiating RL training of a new policy class often presents a different challenge: most deep RL machinery is co-developed with assumptions on the policy class and backbone, resulting in poor performance when the policy class changes. For instance, SAC utilizes a low-variance reparameterization policy gradient for Gaussian policies, but this is unstable for diffusion policies and intractable for autoregressive categorical policies. To address this issue, we develop an offline RL and online fine-tuning approach called policy-agnostic RL (PA-RL) that can effectively train multiple policy classes, with varying architectures and sizes. We build off the basic idea that a universal supervised learning loss can replace the policy improvement step in RL, as long as it is applied on "optimized" actions. To obtain these optimized actions, we first sample multiple actions from a base policy, and run global optimization (i.e., re-ranking multiple action samples using the Q-function) and local optimization (i.e., running gradient steps on an action sample) to maximize the critic on these candidates. PA-RL enables fine-tuning diffusion and transformer policies with either autoregressive tokens or continuous action outputs, at different sizes, entirely via actor-critic RL. Moreover, PA-RL improves the performance and sample-efficiency by up to 2 times compared to existing offline RL and online fine-tuning methods. We show the first result that successfully fine-tunes OpenVLA, a 7B generalist robot policy, autonomously with Cal-QL, an online RL fine-tuning algorithm, improving from 40% to 70% in the real world in 40 minutes.
☆ Toward LLM-Agent-Based Modeling of Transportation Systems: A Conceptual Framework
In transportation system demand modeling and simulation, agent-based models and microsimulations are current state-of-the-art approaches. However, existing agent-based models still have some limitations on behavioral realism and resource demand that limit their applicability. In this study, leveraging the emerging technology of large language models (LLMs) and LLM-based agents, we propose a general LLM-agent-based modeling framework for transportation systems. We argue that LLM agents not only possess the essential capabilities to function as agents but also offer promising solutions to overcome some limitations of existing agent-based models. Our conceptual framework design closely replicates the decision-making and interaction processes and traits of human travelers within transportation networks, and we demonstrate that the proposed systems can meet critical behavioral criteria for decision-making and learning behaviors using related studies and a demonstrative example of LLM agents' learning and adjustment in the bottleneck setting. Although further refinement of the LLM-agent-based modeling framework is necessary, we believe that this approach has the potential to improve transportation system modeling and simulation.
☆ Semantic Search and Recommendation Algorithm
This paper introduces a new semantic search algorithm that uses Word2Vec and Annoy Index to improve the efficiency of information retrieval from large datasets. The proposed approach addresses the limitations of traditional search methods by offering enhanced speed, accuracy, and scalability. Testing on datasets up to 100GB demonstrates the method's effectiveness in processing vast amounts of data while maintaining high precision and performance.
comment: 6 pages, 5 Figures
Detecting Facial Image Manipulations with Multi-Layer CNN Models
The rapid evolution of digital image manipulation techniques poses significant challenges for content verification, with models such as stable diffusion and mid-journey producing highly realistic, yet synthetic, images that can deceive human perception. This research develops and evaluates convolutional neural networks (CNNs) specifically tailored for the detection of these manipulated images. The study implements a comparative analysis of three progressively complex CNN architectures, assessing their ability to classify and localize manipulations across various facial image modifications. Regularization and optimization techniques were systematically incorporated to improve feature extraction and performance. The results indicate that the proposed models achieve an accuracy of up to 76\% in distinguishing manipulated images from genuine ones, surpassing traditional approaches. This research not only highlights the potential of CNNs in enhancing the robustness of digital media verification tools, but also provides insights into effective architectural adaptations and training strategies for low-computation environments. Future work will build on these findings by extending the architectures to handle more diverse manipulation techniques and integrating multi-modal data for improved detection capabilities.
☆ Beyond Scalars: Concept-Based Alignment Analysis in Vision Transformers
Vision transformers (ViTs) can be trained using various learning paradigms, from fully supervised to self-supervised. Diverse training protocols often result in significantly different feature spaces, which are usually compared through alignment analysis. However, current alignment measures quantify this relationship in terms of a single scalar value, obscuring the distinctions between common and unique features in pairs of representations that share the same scalar alignment. We address this limitation by combining alignment analysis with concept discovery, which enables a breakdown of alignment into single concepts encoded in feature space. This fine-grained comparison reveals both universal and unique concepts across different representations, as well as the internal structure of concepts within each of them. Our methodological contributions address two key prerequisites for concept-based alignment: 1) For a description of the representation in terms of concepts that faithfully capture the geometry of the feature space, we define concepts as the most general structure they can possibly form - arbitrary manifolds, allowing hidden features to be described by their proximity to these manifolds. 2) To measure distances between concept proximity scores of two representations, we use a generalized Rand index and partition it for alignment between pairs of concepts. We confirm the superiority of our novel concept definition for alignment analysis over existing linear baselines in a sanity check. The concept-based alignment analysis of representations from four different ViTs reveals that increased supervision correlates with a reduction in the semantic structure of learned representations.
comment: 19 pages, 17 figures, code: https://github.com/jvielhaben/NLMCD-ALIGN
☆ Fundus Image-based Visual Acuity Assessment with PAC-Guarantees ML4H 2024
Timely detection and treatment are essential for maintaining eye health. Visual acuity (VA), which measures the clarity of vision at a distance, is a crucial metric for managing eye health. Machine learning (ML) techniques have been introduced to assist in VA measurement, potentially alleviating clinicians' workloads. However, the inherent uncertainties in ML models make relying solely on them for VA prediction less than ideal. The VA prediction task involves multiple sources of uncertainty, requiring more robust approaches. A promising method is to build prediction sets or intervals rather than point estimates, offering coverage guarantees through techniques like conformal prediction and Probably Approximately Correct (PAC) prediction sets. Despite the potential, to date, these approaches have not been applied to the VA prediction task.To address this, we propose a method for deriving prediction intervals for estimating visual acuity from fundus images with a PAC guarantee. Our experimental results demonstrate that the PAC guarantees are upheld, with performance comparable to or better than that of two prior works that do not provide such guarantees.
comment: To be published in ML4H 2024
☆ Towards Controllable Speech Synthesis in the Era of Large Language Models: A Survey
Text-to-speech (TTS), also known as speech synthesis, is a prominent research area that aims to generate natural-sounding human speech from text. Recently, with the increasing industrial demand, TTS technologies have evolved beyond synthesizing human-like speech to enabling controllable speech generation. This includes fine-grained control over various attributes of synthesized speech such as emotion, prosody, timbre, and duration. Besides, advancements in deep learning, such as diffusion and large language models, have significantly enhanced controllable TTS over the past several years. In this paper, we conduct a comprehensive survey of controllable TTS, covering approaches ranging from basic control techniques to methods utilizing natural language prompts, aiming to provide a clear understanding of the current state of research. We examine the general controllable TTS pipeline, challenges, model architectures, and control strategies, offering a comprehensive and clear taxonomy of existing methods. Additionally, we provide a detailed summary of datasets and evaluation metrics and shed some light on the applications and future directions of controllable TTS. To the best of our knowledge, this survey paper provides the first comprehensive review of emerging controllable TTS methods, which can serve as a beneficial resource for both academic researchers and industry practitioners.
comment: A comprehensive survey on controllable TTS, 23 pages, 6 tables, 4 figures, 280 references
☆ Advancing Music Therapy: Integrating Eastern Five-Element Music Theory and Western Techniques with AI in the Novel Five-Element Harmony System
In traditional medical practices, music therapy has proven effective in treating various psychological and physiological ailments. Particularly in Eastern traditions, the Five Elements Music Therapy (FEMT), rooted in traditional Chinese medicine, possesses profound cultural significance and unique therapeutic philosophies. With the rapid advancement of Information Technology and Artificial Intelligence, applying these modern technologies to FEMT could enhance the personalization and cultural relevance of the therapy and potentially improve therapeutic outcomes. In this article, we developed a music therapy system for the first time by applying the theory of the five elements in music therapy to practice. This innovative approach integrates advanced Information Technology and Artificial Intelligence with Five-Element Music Therapy (FEMT) to enhance personalized music therapy practices. As traditional music therapy predominantly follows Western methodologies, the unique aspects of Eastern practices, specifically the Five-Element theory from traditional Chinese medicine, should be considered. This system aims to bridge this gap by utilizing computational technologies to provide a more personalized, culturally relevant, and therapeutically effective music therapy experience.
comment: 5 pages, 1 figure. Accepted for Publication in the International Symposium on Chinese Spoken Language Processing
☆ EmoSpeech: A Corpus of Emotionally Rich and Contextually Detailed Speech Annotations
Advances in text-to-speech (TTS) technology have significantly improved the quality of generated speech, closely matching the timbre and intonation of the target speaker. However, due to the inherent complexity of human emotional expression, the development of TTS systems capable of controlling subtle emotional differences remains a formidable challenge. Existing emotional speech databases often suffer from overly simplistic labelling schemes that fail to capture a wide range of emotional states, thus limiting the effectiveness of emotion synthesis in TTS applications. To this end, recent efforts have focussed on building databases that use natural language annotations to describe speech emotions. However, these approaches are costly and require more emotional depth to train robust systems. In this paper, we propose a novel process aimed at building databases by systematically extracting emotion-rich speech segments and annotating them with detailed natural language descriptions through a generative model. This approach enhances the emotional granularity of the database and significantly reduces the reliance on costly manual annotations by automatically augmenting the data with high-level language models. The resulting rich database provides a scalable and economically viable solution for developing a more nuanced and dynamic basis for developing emotionally controlled TTS systems.
comment: 4 pages, 1 figure. To appear in the Proceedings of the International Symposium on Chinese Spoken Language Processing, 7-10 November 2024, Beijing, China
☆ ProcessBench: Identifying Process Errors in Mathematical Reasoning
As language models regularly make mistakes when solving math problems, automated identification of errors in the reasoning process becomes increasingly significant for their scalable oversight. In this paper, we introduce ProcessBench for measuring the ability to identify erroneous steps in mathematical reasoning. It consists of 3,400 test cases, primarily focused on competition- and Olympiad-level math problems. Each test case contains a step-by-step solution with error location annotated by human experts. Models are required to identify the earliest step that contains an error, or conclude that all steps are correct. We conduct extensive evaluation on ProcessBench, involving two types of models: process reward models (PRMs) and critic models, where for the latter we prompt general language models to critique each solution step by step. We draw two main observations: (1) Existing PRMs typically fail to generalize to more challenging math problems beyond GSM8K and MATH. They underperform both critic models (i.e., prompted general language models) and our own trained PRM that is straightforwardly fine-tuned on the PRM800K dataset. (2) The best open-source model, QwQ-32B-Preview, has demonstrated the critique capability competitive with the proprietary model GPT-4o, despite that it still lags behind the reasoning-specialized o1-mini. We hope ProcessBench can foster future research in reasoning process assessment, paving the way toward scalable oversight of language models.
☆ Sloth: scaling laws for LLM skills to predict multi-benchmark performance across families
Scaling laws for large language models (LLMs) predict model performance based on parameters like size and training data. However, differences in training configurations and data processing across model families lead to significant variations in benchmark performance, making it difficult for a single scaling law to generalize across all LLMs. On the other hand, training family-specific scaling laws requires training models of varying sizes for every family. In this work, we propose Skills Scaling Laws (SSLaws, pronounced as Sloth), a novel scaling law that leverages publicly available benchmark data and assumes LLM performance is driven by low-dimensional latent skills, such as reasoning and instruction following. These latent skills are influenced by computational resources like model size and training tokens but with varying efficiencies across model families. Sloth exploits correlations across benchmarks to provide more accurate and interpretable predictions while alleviating the need to train multiple LLMs per family. We present both theoretical results on parameter identification and empirical evaluations on 12 prominent benchmarks, from Open LLM Leaderboard v1/v2, demonstrating that Sloth predicts LLM performance efficiently and offers insights into scaling behaviors for downstream tasks such as coding and emotional intelligence applications.
☆ Unraveling the Complexity of Memory in RL Agents: an Approach for Classification and Evaluation
The incorporation of memory into agents is essential for numerous tasks within the domain of Reinforcement Learning (RL). In particular, memory is paramount for tasks that require the utilization of past information, adaptation to novel environments, and improved sample efficiency. However, the term ``memory'' encompasses a wide range of concepts, which, coupled with the lack of a unified methodology for validating an agent's memory, leads to erroneous judgments about agents' memory capabilities and prevents objective comparison with other memory-enhanced agents. This paper aims to streamline the concept of memory in RL by providing practical precise definitions of agent memory types, such as long-term versus short-term memory and declarative versus procedural memory, inspired by cognitive science. Using these definitions, we categorize different classes of agent memory, propose a robust experimental methodology for evaluating the memory capabilities of RL agents, and standardize evaluations. Furthermore, we empirically demonstrate the importance of adhering to the proposed methodology when evaluating different types of agent memory by conducting experiments with different RL agents and what its violation leads to.
comment: 18 pages, 6 figures
☆ HES-UNet: A U-Net for Hepatic Echinococcosis Lesion Segmentation
Hepatic echinococcosis (HE) is a prevalent disease in economically underdeveloped pastoral areas, where adequate medical resources are usually lacking. Existing methods often ignore multi-scale feature fusion or focus only on feature fusion between adjacent levels, which may lead to insufficient feature fusion. To address these issues, we propose HES-UNet, an efficient and accurate model for HE lesion segmentation. This model combines convolutional layers and attention modules to capture local and global features. During downsampling, the multi-directional downsampling block (MDB) is employed to integrate high-frequency and low-frequency features, effectively extracting image details. The multi-scale aggregation block (MAB) aggregates multi-scale feature information. In contrast, the multi-scale upsampling Block (MUB) learns highly abstract features and supplies this information to the skip connection module to fuse multi-scale features. Due to the distinct regional characteristics of HE, there is currently no publicly available high-quality dataset for training our model. We collected CT slice data from 268 patients at a certain hospital to train and evaluate the model. The experimental results show that HES-UNet achieves state-of-the-art performance on our dataset, achieving an overall Dice Similarity Coefficient (DSC) of 89.21%, which is 1.09% higher than that of TransUNet. The project page is available at https://chenjiayan-qhu.github.io/HES-UNet-page.
comment: 6 pages, 3 figures
☆ The Fusion of Large Language Models and Formal Methods for Trustworthy AI Agents: A Roadmap
Large Language Models (LLMs) have emerged as a transformative AI paradigm, profoundly influencing daily life through their exceptional language understanding and contextual generation capabilities. Despite their remarkable performance, LLMs face a critical challenge: the propensity to produce unreliable outputs due to the inherent limitations of their learning-based nature. Formal methods (FMs), on the other hand, are a well-established computation paradigm that provides mathematically rigorous techniques for modeling, specifying, and verifying the correctness of systems. FMs have been extensively applied in mission-critical software engineering, embedded systems, and cybersecurity. However, the primary challenge impeding the deployment of FMs in real-world settings lies in their steep learning curves, the absence of user-friendly interfaces, and issues with efficiency and adaptability. This position paper outlines a roadmap for advancing the next generation of trustworthy AI systems by leveraging the mutual enhancement of LLMs and FMs. First, we illustrate how FMs, including reasoning and certification techniques, can help LLMs generate more reliable and formally certified outputs. Subsequently, we highlight how the advanced learning capabilities and adaptability of LLMs can significantly enhance the usability, efficiency, and scalability of existing FM tools. Finally, we show that unifying these two computation paradigms -- integrating the flexibility and intelligence of LLMs with the rigorous reasoning abilities of FMs -- has transformative potential for the development of trustworthy AI software systems. We acknowledge that this integration has the potential to enhance both the trustworthiness and efficiency of software engineering practices while fostering the development of intelligent FM tools capable of addressing complex yet real-world challenges.
comment: 24 pages, 4 figures
☆ AnomalyControl: Learning Cross-modal Semantic Features for Controllable Anomaly Synthesis
Anomaly synthesis is a crucial approach to augment abnormal data for advancing anomaly inspection. Based on the knowledge from the large-scale pre-training, existing text-to-image anomaly synthesis methods predominantly focus on textual information or coarse-aligned visual features to guide the entire generation process. However, these methods often lack sufficient descriptors to capture the complicated characteristics of realistic anomalies (e.g., the fine-grained visual pattern of anomalies), limiting the realism and generalization of the generation process. To this end, we propose a novel anomaly synthesis framework called AnomalyControl to learn cross-modal semantic features as guidance signals, which could encode the generalized anomaly cues from text-image reference prompts and improve the realism of synthesized abnormal samples. Specifically, AnomalyControl adopts a flexible and non-matching prompt pair (i.e., a text-image reference prompt and a targeted text prompt), where a Cross-modal Semantic Modeling (CSM) module is designed to extract cross-modal semantic features from the textual and visual descriptors. Then, an Anomaly-Semantic Enhanced Attention (ASEA) mechanism is formulated to allow CSM to focus on the specific visual patterns of the anomaly, thus enhancing the realism and contextual relevance of the generated anomaly features. Treating cross-modal semantic features as the prior, a Semantic Guided Adapter (SGA) is designed to encode effective guidance signals for the adequate and controllable synthesis process. Extensive experiments indicate that AnomalyControl can achieve state-of-the-art results in anomaly synthesis compared with existing methods while exhibiting superior performance for downstream tasks.
☆ SimuDICE: Offline Policy Optimization Through World Model Updates and DICE Estimation
In offline reinforcement learning, deriving an effective policy from a pre-collected set of experiences is challenging due to the distribution mismatch between the target policy and the behavioral policy used to collect the data, as well as the limited sample size. Model-based reinforcement learning improves sample efficiency by generating simulated experiences using a learned dynamic model of the environment. However, these synthetic experiences often suffer from the same distribution mismatch. To address these challenges, we introduce SimuDICE, a framework that iteratively refines the initial policy derived from offline data using synthetically generated experiences from the world model. SimuDICE enhances the quality of these simulated experiences by adjusting the sampling probabilities of state-action pairs based on stationary DIstribution Correction Estimation (DICE) and the estimated confidence in the model's predictions. This approach guides policy improvement by balancing experiences similar to those frequently encountered with ones that have a distribution mismatch. Our experiments show that SimuDICE achieves performance comparable to existing algorithms while requiring fewer pre-collected experiences and planning steps, and it remains robust across varying data collection policies.
comment: Published at BNAIC/BeNeLearn 2024
☆ SafeWorld: Geo-Diverse Safety Alignment NeurIPS 2024
In the rapidly evolving field of Large Language Models (LLMs), ensuring safety is a crucial and widely discussed topic. However, existing works often overlook the geo-diversity of cultural and legal standards across the world. To demonstrate the challenges posed by geo-diverse safety standards, we introduce SafeWorld, a novel benchmark specifically designed to evaluate LLMs' ability to generate responses that are not only helpful but also culturally sensitive and legally compliant across diverse global contexts. SafeWorld encompasses 2,342 test user queries, each grounded in high-quality, human-verified cultural norms and legal policies from 50 countries and 493 regions/races. On top of it, we propose a multi-dimensional automatic safety evaluation framework that assesses the contextual appropriateness, accuracy, and comprehensiveness of responses. Our evaluations reveal that current LLMs struggle to meet these criteria. To enhance LLMs' alignment with geo-diverse safety standards, we synthesize helpful preference pairs for Direct Preference Optimization (DPO) alignment training. The preference pair construction aims to encourage LLMs to behave appropriately and provide precise references to relevant cultural norms and policies when necessary. Our trained SafeWorldLM outperforms all competing models, including GPT-4o on all three evaluation dimensions by a large margin. Global human evaluators also note a nearly 20% higher winning rate in helpfulness and harmfulness evaluation. Our code and data can be found here: https://github.com/PlusLabNLP/SafeWorld.
comment: Accepted by NeurIPS 2024
☆ From Uncertainty to Trust: Enhancing Reliability in Vision-Language Models with Uncertainty-Guided Dropout Decoding
Large vision-language models (LVLMs) demonstrate remarkable capabilities in multimodal tasks but are prone to misinterpreting visual inputs, often resulting in hallucinations and unreliable outputs. To address these challenges, we propose Dropout Decoding, a novel inference-time approach that quantifies the uncertainty of visual tokens and selectively masks uncertain tokens to improve decoding. Our method measures the uncertainty of each visual token by projecting it onto the text space and decomposing it into aleatoric and epistemic components. Specifically, we focus on epistemic uncertainty, which captures perception-related errors more effectively. Inspired by dropout regularization, we introduce uncertainty-guided token dropout, which applies the dropout principle to input visual tokens instead of model parameters, and during inference rather than training. By aggregating predictions from an ensemble of masked decoding contexts, Dropout Decoding robustly mitigates errors arising from visual token misinterpretations. Evaluations on benchmarks including CHAIR, THRONE, and MMBench demonstrate that Dropout Decoding significantly reduces object hallucinations (OH) and enhances both reliability and quality of LVLM outputs across diverse visual contexts.
comment: Code is released at https://github.com/kigb/DropoutDecoding
☆ How Certain are Uncertainty Estimates? Three Novel Earth Observation Datasets for Benchmarking Uncertainty Quantification in Machine Learning
Uncertainty quantification (UQ) is essential for assessing the reliability of Earth observation (EO) products. However, the extensive use of machine learning models in EO introduces an additional layer of complexity, as those models themselves are inherently uncertain. While various UQ methods do exist for machine learning models, their performance on EO datasets remains largely unevaluated. A key challenge in the community is the absence of the ground truth for uncertainty, i.e. how certain the uncertainty estimates are, apart from the labels for the image/signal. This article fills this gap by introducing three benchmark datasets specifically designed for UQ in EO machine learning models. These datasets address three common problem types in EO: regression, image segmentation, and scene classification. They enable a transparent comparison of different UQ methods for EO machine learning models. We describe the creation and characteristics of each dataset, including data sources, preprocessing steps, and label generation, with a particular focus on calculating the reference uncertainty. We also showcase baseline performance of several machine learning models on each dataset, highlighting the utility of these benchmarks for model development and comparison. Overall, this article offers a valuable resource for researchers and practitioners working in artificial intelligence for EO, promoting a more accurate and reliable quality measure of the outputs of machine learning models. The dataset and code are accessible via https://gitlab.lrz.de/ai4eo/WG_Uncertainty.
comment: Submitted to IEEE Geoscience and Remote Sensing Magazine
☆ Simulating Human-like Daily Activities with Desire-driven Autonomy
Existing task-oriented AI agents often depend on explicit instructions or external rewards, limiting their ability to be driven by intrinsic motivations like humans. In this paper, we present a desire-driven autonomy framework to guide a Large Language Model-based (LLM-based) agent to simulate human-like daily activities. In contrast to previous agents, our Desire-driven Autonomous Agent (D2A) operates on the principle of intrinsic desire, allowing it to propose and select tasks that fulfill its motivational framework autonomously. Inspired by the Theory of Needs, the motivational framework incorporates an understanding of human-like desires, such as the need for social interaction, personal fulfillment, and self-care. Utilizing a desire-driven task generation mechanism, the agent evaluates its current state and takes a sequence of activities aligned with its intrinsic motivations. Through simulations, we demonstrate that our Desire-driven Autonomous Agent (D2A) generates coherent, contextually relevant daily activities while exhibiting variability and adaptability similar to human behavior. A comparative analysis with other LLM-based frameworks demonstrates that our approach significantly enhances the rationality of the simulated activities.
LLM-BIP: Structured Pruning for Large Language Models with Block-Wise Forward Importance Propagation
Large language models (LLMs) have demonstrated remarkable performance across various language tasks, but their widespread deployment is impeded by their large size and high computational costs. Structural pruning is a prevailing technique used to introduce sparsity into pre-trained models and facilitate direct hardware acceleration during inference by removing redundant connections (structurally-grouped parameters), such as channels and attention heads. Existing structural pruning approaches often employ either global or layer-wise pruning criteria; however, they are hindered by ineffectiveness stemming from inaccurate evaluation of connection importance. Global pruning methods typically assess component importance using near-zero and unreliable gradients, while layer-wise pruning approaches encounter significant pruning error accumulation issues. To this end, we propose a more accurate pruning metric based on the block-wise importance score propagation, termed LLM-BIP. Specifically, LLM-BIP precisely evaluates connection importance by gauging its influence on the respective transformer block output, which can be efficiently approximated in a single forward pass through an upper bound derived from the assumption of Lipschitz continuity. We evaluate the proposed method using LLaMA-7B, Vicuna-7B, and LLaMA-13B across common zero-shot tasks. The results demonstrate that our approach achieves an average of 3.26% increase in accuracy for common reasoning tasks compared to previous best baselines. It also reduces perplexity by 14.09 and 68.76 on average for the WikiText2 dataset and PTB dataset, respectively.
☆ StarWhisper Telescope: Agent-Based Observation Assistant System to Approach AI Astrophysicist
With the rapid advancements in Large Language Models (LLMs), LLM-based agents have introduced convenient and user-friendly methods for leveraging tools across various domains. In the field of astronomical observation, the construction of new telescopes has significantly increased astronomers' workload. Deploying LLM-powered agents can effectively alleviate this burden and reduce the costs associated with training personnel. Within the Nearby Galaxy Supernovae Survey (NGSS) project, which encompasses eight telescopes across three observation sites, aiming to find the transients from the galaxies in 50 mpc, we have developed the \textbf{StarWhisper Telescope System} to manage the entire observation process. This system automates tasks such as generating observation lists, conducting observations, analyzing data, and providing feedback to the observer. Observation lists are customized for different sites and strategies to ensure comprehensive coverage of celestial objects. After manual verification, these lists are uploaded to the telescopes via the agents in the system, which initiates observations upon neutral language. The observed images are analyzed in real-time, and the transients are promptly communicated to the observer. The agent modifies them into a real-time follow-up observation proposal and send to the Xinglong observatory group chat, then add them to the next-day observation lists. Additionally, the integration of AI agents within the system provides online accessibility, saving astronomers' time and encouraging greater participation from amateur astronomers in the NGSS project.
comment: 21 pages, 18 figures
☆ BatchTopK Sparse Autoencoders
Sparse autoencoders (SAEs) have emerged as a powerful tool for interpreting language model activations by decomposing them into sparse, interpretable features. A popular approach is the TopK SAE, that uses a fixed number of the most active latents per sample to reconstruct the model activations. We introduce BatchTopK SAEs, a training method that improves upon TopK SAEs by relaxing the top-k constraint to the batch-level, allowing for a variable number of latents to be active per sample. As a result, BatchTopK adaptively allocates more or fewer latents depending on the sample, improving reconstruction without sacrificing average sparsity. We show that BatchTopK SAEs consistently outperform TopK SAEs in reconstructing activations from GPT-2 Small and Gemma 2 2B, and achieve comparable performance to state-of-the-art JumpReLU SAEs. However, an advantage of BatchTopK is that the average number of latents can be directly specified, rather than approximately tuned through a costly hyperparameter sweep. We provide code for training and evaluating BatchTopK SAEs at https://github.com/bartbussmann/BatchTopK
☆ GameArena: Evaluating LLM Reasoning through Live Computer Games
Evaluating the reasoning abilities of large language models (LLMs) is challenging. Existing benchmarks often depend on static datasets, which are vulnerable to data contamination and may get saturated over time, or on binary live human feedback that conflates reasoning with other abilities. As the most prominent dynamic benchmark, Chatbot Arena evaluates open-ended questions in real-world settings, but lacks the granularity in assessing specific reasoning capabilities. We introduce GameArena, a dynamic benchmark designed to evaluate LLM reasoning capabilities through interactive gameplay with humans. GameArena consists of three games designed to test specific reasoning capabilities (e.g., deductive and inductive reasoning), while keeping participants entertained and engaged. We analyze the gaming data retrospectively to uncover the underlying reasoning processes of LLMs and measure their fine-grained reasoning capabilities. We collect over 2000 game sessions and provide detailed assessments of various reasoning capabilities for five state-of-the-art LLMs. Our user study with 100 participants suggests that GameArena improves user engagement compared to Chatbot Arena. For the first time, GameArena enables the collection of step-by-step LLM reasoning data in the wild.
☆ Edge Delayed Deep Deterministic Policy Gradient: efficient continuous control for edge scenarios
Deep Reinforcement Learning is gaining increasing attention thanks to its capability to learn complex policies in high-dimensional settings. Recent advancements utilize a dual-network architecture to learn optimal policies through the Q-learning algorithm. However, this approach has notable drawbacks, such as an overestimation bias that can disrupt the learning process and degrade the performance of the resulting policy. To address this, novel algorithms have been developed that mitigate overestimation bias by employing multiple Q-functions. Edge scenarios, which prioritize privacy, have recently gained prominence. In these settings, limited computational resources pose a significant challenge for complex Machine Learning approaches, making the efficiency of algorithms crucial for their performance. In this work, we introduce a novel Reinforcement Learning algorithm tailored for edge scenarios, called Edge Delayed Deep Deterministic Policy Gradient (EdgeD3). EdgeD3 enhances the Deep Deterministic Policy Gradient (DDPG) algorithm, achieving significantly improved performance with $25\%$ less Graphics Process Unit (GPU) time while maintaining the same memory usage. Additionally, EdgeD3 consistently matches or surpasses the performance of state-of-the-art methods across various benchmarks, all while using $30\%$ fewer computational resources and requiring $30\%$ less memory.
☆ Exploring Memorization and Copyright Violation in Frontier LLMs: A Study of the New York Times v. OpenAI 2023 Lawsuit
Copyright infringement in frontier LLMs has received much attention recently due to the New York Times v. OpenAI lawsuit, filed in December 2023. The New York Times claims that GPT-4 has infringed its copyrights by reproducing articles for use in LLM training and by memorizing the inputs, thereby publicly displaying them in LLM outputs. Our work aims to measure the propensity of OpenAI's LLMs to exhibit verbatim memorization in its outputs relative to other LLMs, specifically focusing on news articles. We discover that both GPT and Claude models use refusal training and output filters to prevent verbatim output of the memorized articles. We apply a basic prompt template to bypass the refusal training and show that OpenAI models are currently less prone to memorization elicitation than models from Meta, Mistral, and Anthropic. We find that as models increase in size, especially beyond 100 billion parameters, they demonstrate significantly greater capacity for memorization. Our findings have practical implications for training: more attention must be placed on preventing verbatim memorization in very large models. Our findings also have legal significance: in assessing the relative memorization capacity of OpenAI's LLMs, we probe the strength of The New York Times's copyright infringement claims and OpenAI's legal defenses, while underscoring issues at the intersection of generative AI, law, and policy.
☆ Measuring Pre-training Data Quality without Labels for Time Series Foundation Models
Recently, there has been a growing interest in time series foundation models that generalize across different downstream tasks. A key to strong foundation models is a diverse pre-training dataset, which is particularly challenging to collect for time series classification. In this work, we explore the performance of a contrastive-learning-based foundation model as a function of the data used for pre-training. We introduce contrastive accuracy, a new measure to evaluate the quality of the representation space learned by the foundation model. Our experiments reveal the positive correlation between the proposed measure and the accuracy of the model on a collection of downstream tasks. This suggests that the contrastive accuracy can serve as a criterion to search for time series datasets that can enhance the pre-training and improve thereby the foundation model's generalization.
☆ Elastic-DETR: Making Image Resolution Learnable with Content-Specific Network Prediction
Multi-scale image resolution is a de facto standard approach in modern object detectors, such as DETR. This technique allows for the acquisition of various scale information from multiple image resolutions. However, manual hyperparameter selection of the resolution can restrict its flexibility, which is informed by prior knowledge, necessitating human intervention. This work introduces a novel strategy for learnable resolution, called Elastic-DETR, enabling elastic utilization of multiple image resolutions. Our network provides an adaptive scale factor based on the content of the image with a compact scale prediction module (< 2 GFLOPs). The key aspect of our method lies in how to determine the resolution without prior knowledge. We present two loss functions derived from identified key components for resolution optimization: scale loss, which increases adaptiveness according to the image, and distribution loss, which determines the overall degree of scaling based on network performance. By leveraging the resolution's flexibility, we can demonstrate various models that exhibit varying trade-offs between accuracy and computational complexity. We empirically show that our scheme can unleash the potential of a wide spectrum of image resolutions without constraining flexibility. Our models on MS COCO establish a maximum accuracy gain of 3.5%p or 26% decrease in computation than MS-trained DN-DETR.
☆ Augmenting the action space with conventions to improve multi-agent cooperation in Hanabi AAMAS
The card game Hanabi is considered a strong medium for the testing and development of multi-agent reinforcement learning (MARL) algorithms, due to its cooperative nature, hidden information, limited communication and remarkable complexity. Previous research efforts have explored the capabilities of MARL algorithms within Hanabi, focusing largely on advanced architecture design and algorithmic manipulations to achieve state-of-the-art performance for a various number of cooperators. However, this often leads to complex solution strategies with high computational cost and requiring large amounts of training data. For humans to solve the Hanabi game effectively, they require the use of conventions, which often allows for a means to implicitly convey ideas or knowledge based on a predefined, and mutually agreed upon, set of ``rules''. Multi-agent problems containing partial observability, especially when limited communication is present, can benefit greatly from the use of implicit knowledge sharing. In this paper, we propose a novel approach to augmenting the action space using conventions, which act as special cooperative actions that span over multiple time steps and multiple agents, requiring agents to actively opt in for it to reach fruition. These conventions are based on existing human conventions, and result in a significant improvement on the performance of existing techniques for self-play and cross-play across a various number of cooperators within Hanabi.
comment: This paper is under review at the journal of autonomous agents and multi-agent systems (JAAMAS)
☆ Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection SC
Automatic Speech Recognition (ASR) plays an important role in speech-based automatic detection of Alzheimer's disease (AD). However, recognition errors could propagate downstream, potentially impacting the detection decisions. Recent studies have revealed a non-linear relationship between word error rates (WER) and AD detection performance, where ASR transcriptions with notable errors could still yield AD detection accuracy equivalent to that based on manual transcriptions. This work presents a series of analyses to explore the effect of ASR transcription errors in BERT-based AD detection systems. Our investigation reveals that not all ASR errors contribute equally to detection performance. Certain words, such as stopwords, despite constituting a large proportion of errors, are shown to play a limited role in distinguishing AD. In contrast, the keywords related to diagnosis tasks exhibit significantly greater importance relative to other words. These findings provide insights into the interplay between ASR errors and the downstream detection model.
comment: Accepted by IEEE ISCSLP 2024
☆ CAD-Unet: A Capsule Network-Enhanced Unet Architecture for Accurate Segmentation of COVID-19 Lung Infections from CT Images
Since the outbreak of the COVID-19 pandemic in 2019, medical imaging has emerged as a primary modality for diagnosing COVID-19 pneumonia. In clinical settings, the segmentation of lung infections from computed tomography images enables rapid and accurate quantification and diagnosis of COVID-19. Segmentation of COVID-19 infections in the lungs poses a formidable challenge, primarily due to the indistinct boundaries and limited contrast presented by ground glass opacity manifestations. Moreover, the confounding similarity between infiltrates, lung tissues, and lung walls further complicates this segmentation task. To address these challenges, this paper introduces a novel deep network architecture, called CAD-Unet, for segmenting COVID-19 lung infections. In this architecture, capsule networks are incorporated into the existing Unet framework. Capsule networks represent a novel network architecture that differs from traditional convolutional neural networks. They utilize vectors for information transfer among capsules, facilitating the extraction of intricate lesion spatial information. Additionally, we design a capsule encoder path and establish a coupling path between the unet encoder and the capsule encoder. This design maximizes the complementary advantages of both network structures while achieving efficient information fusion. \noindent Finally, extensive experiments are conducted on four publicly available datasets, encompassing binary segmentation tasks and multi-class segmentation tasks. The experimental results demonstrate the superior segmentation performance of the proposed model. The code has been released at: https://github.com/AmanoTooko-jie/CAD-Unet.
☆ Towards High-Level Modelling in Automated Planning
Planning is a fundamental activity, arising frequently in many contexts, from daily tasks to industrial processes. The planning task consists of selecting a sequence of actions to achieve a specified goal from specified initial conditions. The Planning Domain Definition Language (PDDL) is the leading language used in the field of automated planning to model planning problems. Previous work has highlighted the limitations of PDDL, particularly in terms of its expressivity. Our interest lies in facilitating the handling of complex problems and enhancing the overall capability of automated planning systems. Unified-Planning is a Python library offering high-level API to specify planning problems and to invoke automated planners. In this paper, we present an extension of the UP library aimed at enhancing its expressivity for high-level problem modelling. In particular, we have added an array type, an expression to count booleans, and the allowance for integer parameters in actions. We show how these facilities enable natural high-level models of three classical planning problems.
☆ PRECISE: Pre-training Sequential Recommenders with Collaborative and Semantic Information
Real-world recommendation systems commonly offer diverse content scenarios for users to interact with. Considering the enormous number of users in industrial platforms, it is infeasible to utilize a single unified recommendation model to meet the requirements of all scenarios. Usually, separate recommendation pipelines are established for each distinct scenario. This practice leads to challenges in comprehensively grasping users' interests. Recent research endeavors have been made to tackle this problem by pre-training models to encapsulate the overall interests of users. Traditional pre-trained recommendation models mainly capture user interests by leveraging collaborative signals. Nevertheless, a prevalent drawback of these systems is their incapacity to handle long-tail items and cold-start scenarios. With the recent advent of large language models, there has been a significant increase in research efforts focused on exploiting LLMs to extract semantic information for users and items. However, text-based recommendations highly rely on elaborate feature engineering and frequently fail to capture collaborative similarities. To overcome these limitations, we propose a novel pre-training framework for sequential recommendation, termed PRECISE. This framework combines collaborative signals with semantic information. Moreover, PRECISE employs a learning framework that initially models users' comprehensive interests across all recommendation scenarios and subsequently concentrates on the specific interests of target-scene behaviors. We demonstrate that PRECISE precisely captures the entire range of user interests and effectively transfers them to the target interests. Empirical findings reveal that the PRECISE framework attains outstanding performance on both public and industrial datasets.
☆ DSAI: Unbiased and Interpretable Latent Feature Extraction for Data-Centric AI
Large language models (LLMs) often struggle to objectively identify latent characteristics in large datasets due to their reliance on pre-trained knowledge rather than actual data patterns. To address this data grounding issue, we propose Data Scientist AI (DSAI), a framework that enables unbiased and interpretable feature extraction through a multi-stage pipeline with quantifiable prominence metrics for evaluating extracted features. On synthetic datasets with known ground-truth features, DSAI demonstrates high recall in identifying expert-defined features while faithfully reflecting the underlying data. Applications on real-world datasets illustrate the framework's practical utility in uncovering meaningful patterns with minimal expert oversight, supporting use cases such as interpretable classification. The title of our paper is chosen from multiple candidates based on DSAI-generated criteria.
☆ S$^{2}$FT: Efficient, Scalable and Generalizable LLM Fine-tuning by Structured Sparsity
Current PEFT methods for LLMs can achieve either high quality, efficient training, or scalable serving, but not all three simultaneously. To address this limitation, we investigate sparse fine-tuning and observe a remarkable improvement in generalization ability. Utilizing this key insight, we propose a family of Structured Sparse Fine-Tuning (S$^{2}$FT) methods for LLMs, which concurrently achieve state-of-the-art fine-tuning performance, training efficiency, and inference scalability. S$^{2}$FT accomplishes this by "selecting sparsely and computing densely". It selects a few heads and channels in the MHA and FFN modules for each Transformer block, respectively. Next, it co-permutes weight matrices on both sides of the coupled structures in LLMs to connect the selected components in each layer into a dense submatrix. Finally, S$^{2}$FT performs in-place gradient updates on all submatrices. Through theoretical analysis and empirical results, our method prevents overfitting and forgetting, delivers SOTA performance on both commonsense and arithmetic reasoning with 4.6% and 1.3% average improvements compared to LoRA, and surpasses full FT by 11.5% when generalizing to various domains after instruction tuning. Using our partial backpropagation algorithm, S$^{2}$FT saves training memory up to 3$\times$ and improves latency by 1.5-2.7$\times$ compared to full FT, while delivering an average 10% improvement over LoRA on both metrics. We further demonstrate that the weight updates in S$^{2}$FT can be decoupled into adapters, enabling effective fusion, fast switch, and efficient parallelism for serving multiple fine-tuned models.
☆ Methods for Legal Citation Prediction in the Age of LLMs: An Australian Law Case Study
In recent years, Large Language Models (LLMs) have shown great potential across a wide range of legal tasks. Despite these advances, mitigating hallucination remains a significant challenge, with state-of-the-art LLMs still frequently generating incorrect legal references. In this paper, we focus on the problem of legal citation prediction within the Australian law context, where correctly identifying and citing relevant legislations or precedents is critical. We compare several approaches: prompting general purpose and law-specialised LLMs, retrieval-only pipelines with both generic and domain-specific embeddings, task-specific instruction-tuning of LLMs, and hybrid strategies that combine LLMs with retrieval augmentation, query expansion, or voting ensembles. Our findings indicate that domain-specific pre-training alone is insufficient for achieving satisfactory citation accuracy even after law-specialised pre-training. In contrast, instruction tuning on our task-specific dataset dramatically boosts performance reaching the best results across all settings. We also highlight that database granularity along with the type of embeddings play a critical role in the performance of retrieval systems. Among retrieval-based approaches, hybrid methods consistently outperform retrieval-only setups, and among these, ensemble voting delivers the best result by combining the predictive quality of instruction-tuned LLMs with the retrieval system.
comment: For code, data, and models see https://auslawbench.github.io
☆ A Lightweight U-like Network Utilizing Neural Memory Ordinary Differential Equations for Slimming the Decoder
In recent years, advanced U-like networks have demonstrated remarkable performance in medical image segmentation tasks. However, their drawbacks, including excessive parameters, high computational complexity, and slow inference speed, pose challenges for practical implementation in scenarios with limited computational resources. Existing lightweight U-like networks have alleviated some of these problems, but they often have pre-designed structures and consist of inseparable modules, limiting their application scenarios. In this paper, we propose three plug-and-play decoders by employing different discretization methods of the neural memory Ordinary Differential Equations (nmODEs). These decoders integrate features at various levels of abstraction by processing information from skip connections and performing numerical operations on upward path. Through experiments on the PH2, ISIC2017, and ISIC2018 datasets, we embed these decoders into different U-like networks, demonstrating their effectiveness in significantly reducing the number of parameters and FLOPs while maintaining performance. In summary, the proposed discretized nmODEs decoders are capable of reducing the number of parameters by about 20% ~ 50% and FLOPs by up to 74%, while possessing the potential to adapt to all U-like networks. Our code is available at https://github.com/nayutayuki/Lightweight-nmODE-Decoders-For-U-like-networks.
☆ Unseen Attack Detection in Software-Defined Networking Using a BERT-Based Large Language Model
Software defined networking (SDN) represents a transformative shift in network architecture by decoupling the control plane from the data plane, enabling centralized and flexible management of network resources. However, this architectural shift introduces significant security challenges, as SDN's centralized control becomes an attractive target for various types of attacks. While current research has yielded valuable insights into attack detection in SDN, critical gaps remain. Addressing challenges in feature selection, broadening the scope beyond DDoS attacks, strengthening attack decisions based on multi flow analysis, and building models capable of detecting unseen attacks that they have not been explicitly trained on are essential steps toward advancing security in SDN. In this paper, we introduce a novel approach that leverages Natural Language Processing (NLP) and the pre trained BERT base model to enhance attack detection in SDN. Our approach transforms network flow data into a format interpretable by language models, allowing BERT to capture intricate patterns and relationships within network traffic. By using Random Forest for feature selection, we optimize model performance and reduce computational overhead, ensuring accurate detection. Attack decisions are made based on several flows, providing stronger and more reliable detection of malicious traffic. Furthermore, our approach is specifically designed to detect previously unseen attacks, offering a solution for identifying threats that the model was not explicitly trained on. To rigorously evaluate our approach, we conducted experiments in two scenarios: one focused on detecting known attacks, achieving 99.96% accuracy, and another on detecting unseen attacks, where our model achieved 99.96% accuracy, demonstrating the robustness of our approach in detecting evolving threats to improve the security of SDN networks.
comment: Mohammed N. Swileh is first author. Shengli Zhang is corresponding author
LLMs as Debate Partners: Utilizing Genetic Algorithms and Adversarial Search for Adaptive Arguments
This paper introduces DebateBrawl, an innovative AI-powered debate platform that integrates Large Language Models (LLMs), Genetic Algorithms (GA), and Adversarial Search (AS) to create an adaptive and engaging debating experience. DebateBrawl addresses the limitations of traditional LLMs in strategic planning by incorporating evolutionary optimization and game-theoretic techniques. The system demonstrates remarkable performance in generating coherent, contextually relevant arguments while adapting its strategy in real-time. Experimental results involving 23 debates show balanced outcomes between AI and human participants, with the AI system achieving an average score of 2.72 compared to the human average of 2.67 out of 10. User feedback indicates significant improvements in debating skills and a highly satisfactory learning experience, with 85% of users reporting improved debating abilities and 78% finding the AI opponent appropriately challenging. The system's ability to maintain high factual accuracy (92% compared to 78% in human-only debates) while generating diverse arguments addresses critical concerns in AI-assisted discourse. DebateBrawl not only serves as an effective educational tool but also contributes to the broader goal of improving public discourse through AI-assisted argumentation. The paper discusses the ethical implications of AI in persuasive contexts and outlines the measures implemented to ensure responsible development and deployment of the system, including robust fact-checking mechanisms and transparency in decision-making processes.
☆ Data Free Backdoor Attacks NeurIPS 2024
Backdoor attacks aim to inject a backdoor into a classifier such that it predicts any input with an attacker-chosen backdoor trigger as an attacker-chosen target class. Existing backdoor attacks require either retraining the classifier with some clean data or modifying the model's architecture. As a result, they are 1) not applicable when clean data is unavailable, 2) less efficient when the model is large, and 3) less stealthy due to architecture changes. In this work, we propose DFBA, a novel retraining-free and data-free backdoor attack without changing the model architecture. Technically, our proposed method modifies a few parameters of a classifier to inject a backdoor. Through theoretical analysis, we verify that our injected backdoor is provably undetectable and unremovable by various state-of-the-art defenses under mild assumptions. Our evaluation on multiple datasets further demonstrates that our injected backdoor: 1) incurs negligible classification loss, 2) achieves 100% attack success rates, and 3) bypasses six existing state-of-the-art defenses. Moreover, our comparison with a state-of-the-art non-data-free backdoor attack shows our attack is more stealthy and effective against various defenses while achieving less classification accuracy loss.
comment: 24 pages, 8 figures, accepted by NeurIPS 2024
☆ A Real-Time Defense Against Object Vanishing Adversarial Patch Attacks for Object Detection in Autonomous Vehicles
Autonomous vehicles (AVs) increasingly use DNN-based object detection models in vision-based perception. Correct detection and classification of obstacles is critical to ensure safe, trustworthy driving decisions. Adversarial patches aim to fool a DNN with intentionally generated patterns concentrated in a localized region of an image. In particular, object vanishing patch attacks can cause object detection models to fail to detect most or all objects in a scene, posing a significant practical threat to AVs. This work proposes ADAV (Adversarial Defense for Autonomous Vehicles), a novel defense methodology against object vanishing patch attacks specifically designed for autonomous vehicles. Unlike existing defense methods which have high latency or are designed for static images, ADAV runs in real-time and leverages contextual information from prior frames in an AV's video feed. ADAV checks if the object detector's output for the target frame is temporally consistent with the output from a previous reference frame to detect the presence of a patch. If the presence of a patch is detected, ADAV uses gradient-based attribution to localize adversarial pixels that break temporal consistency. This two stage procedure allows ADAV to efficiently process clean inputs, and both stages are optimized to be low latency. ADAV is evaluated using real-world driving data from the Berkeley Deep Drive BDD100K dataset, and demonstrates high adversarial and clean performance.
☆ A Self-guided Multimodal Approach to Enhancing Graph Representation Learning for Alzheimer's Diseases
Graph neural networks (GNNs) are powerful machine learning models designed to handle irregularly structured data. However, their generic design often proves inadequate for analyzing brain connectomes in Alzheimer's Disease (AD), highlighting the need to incorporate domain knowledge for optimal performance. Infusing AD-related knowledge into GNNs is a complicated task. Existing methods typically rely on collaboration between computer scientists and domain experts, which can be both time-intensive and resource-demanding. To address these limitations, this paper presents a novel self-guided, knowledge-infused multimodal GNN that autonomously incorporates domain knowledge into the model development process. Our approach conceptualizes domain knowledge as natural language and introduces a specialized multimodal GNN capable of leveraging this uncurated knowledge to guide the learning process of the GNN, such that it can improve the model performance and strengthen the interpretability of the predictions. To evaluate our framework, we curated a comprehensive dataset of recent peer-reviewed papers on AD and integrated it with multiple real-world AD datasets. Experimental results demonstrate the ability of our method to extract relevant domain knowledge, provide graph-based explanations for AD diagnosis, and improve the overall performance of the GNN. This approach provides a more scalable and efficient alternative to inject domain knowledge for AD compared with the manual design from the domain expert, advancing both prediction accuracy and interpretability in AD diagnosis.
☆ MSCrackMamba: Leveraging Vision Mamba for Crack Detection in Fused Multispectral Imagery
Crack detection is a critical task in structural health monitoring, aimed at assessing the structural integrity of bridges, buildings, and roads to prevent potential failures. Vision-based crack detection has become the mainstream approach due to its ease of implementation and effectiveness. Fusing infrared (IR) channels with red, green and blue (RGB) channels can enhance feature representation and thus improve crack detection. However, IR and RGB channels often differ in resolution. To align them, higher-resolution RGB images typically need to be downsampled to match the IR image resolution, which leads to the loss of fine details. Moreover, crack detection performance is restricted by the limited receptive fields and high computational complexity of traditional image segmentation networks. Inspired by the recently proposed Mamba neural architecture, this study introduces a two-stage paradigm called MSCrackMamba, which leverages Vision Mamba along with a super-resolution network to address these challenges. Specifically, to align IR and RGB channels, we first apply super-resolution to IR channels to match the resolution of RGB channels for data fusion. Vision Mamba is then adopted as the backbone network, while UperNet is employed as the decoder for crack detection. Our approach is validated on the large-scale Crack Detection dataset Crack900, demonstrating an improvement of 3.55% in mIoU compared to the best-performing baseline methods.
☆ Skill-Enhanced Reinforcement Learning Acceleration from Demonstrations ICML 2024
Learning from Demonstration (LfD) aims to facilitate rapid Reinforcement Learning (RL) by leveraging expert demonstrations to pre-train the RL agent. However, the limited availability of expert demonstration data often hinders its ability to effectively aid downstream RL learning. To address this problem, we propose a novel two-stage method dubbed as Skill-enhanced Reinforcement Learning Acceleration (SeRLA). SeRLA introduces a skill-level adversarial Positive-Unlabeled (PU) learning model to extract useful skill prior knowledge by enabling learning from both limited expert data and general low-cost demonstration data in the offline prior learning stage. Subsequently, it deploys a skill-based soft actor-critic algorithm to leverage this acquired prior knowledge in the downstream online RL stage for efficient training of a skill policy network. Moreover, we develop a simple skill-level data enhancement technique to further alleviate data sparsity and improve both skill prior learning and downstream skill policy training. Our experimental results on multiple standard RL environments show the proposed SeRLA method achieves state-of-the-art performance on accelerating reinforcement learning on downstream tasks, especially in the early learning phase.
comment: ICML 2024 AutoRL Workshop; 9 pages
☆ Enhancing Adversarial Resistance in LLMs with Recursion
The increasing integration of Large Language Models (LLMs) into society necessitates robust defenses against vulnerabilities from jailbreaking and adversarial prompts. This project proposes a recursive framework for enhancing the resistance of LLMs to manipulation through the use of prompt simplification techniques. By increasing the transparency of complex and confusing adversarial prompts, the proposed method enables more reliable detection and prevention of malicious inputs. Our findings attempt to address a critical problem in AI safety and security, providing a foundation for the development of systems able to distinguish harmless inputs from prompts containing malicious intent. As LLMs continue to be used in diverse applications, the importance of such safeguards will only grow.
☆ Annotations for Exploring Food Tweets From Multiple Aspects
This research builds upon the Latvian Twitter Eater Corpus (LTEC), which is focused on the narrow domain of tweets related to food, drinks, eating and drinking. LTEC has been collected for more than 12 years and reaching almost 3 million tweets with the basic information as well as extended automatically and manually annotated metadata. In this paper we supplement the LTEC with manually annotated subsets of evaluation data for machine translation, named entity recognition, timeline-balanced sentiment analysis, and text-image relation classification. We experiment with each of the data sets using baseline models and highlight future challenges for various modelling approaches.
☆ AlphaVerus: Bootstrapping Formally Verified Code Generation through Self-Improving Translation and Treefinement
Automated code generation with large language models has gained significant traction, but there remains no guarantee on the correctness of generated code. We aim to use formal verification to provide mathematical guarantees that the generated code is correct. However, generating formally verified code with LLMs is hindered by the scarcity of training data and the complexity of formal proofs. To tackle this challenge, we introduce AlphaVerus, a self-improving framework that bootstraps formally verified code generation by iteratively translating programs from a higher-resource language and leveraging feedback from a verifier. AlphaVerus operates in three phases: exploration of candidate translations, Treefinement -- a novel tree search algorithm for program refinement using verifier feedback, and filtering misaligned specifications and programs to prevent reward hacking. Through this iterative process, AlphaVerus enables a LLaMA-3.1-70B model to generate verified code without human intervention or model finetuning. AlphaVerus shows an ability to generate formally verified solutions for HumanEval and MBPP, laying the groundwork for truly trustworthy code-generation agents.
☆ ACQ: A Unified Framework for Automated Programmatic Creativity in Online Advertising
In online advertising, the demand-side platform (a.k.a. DSP) enables advertisers to create different ad creatives for real-time bidding. Intuitively, advertisers tend to create more ad creatives for a single photo to increase the probability of participating in bidding, further enhancing their ad cost. From the perspective of DSP, the following are two overlooked issues. On the one hand, the number of ad creatives cannot grow indefinitely. On the other hand, the marginal effects of ad cost diminish as the number of ad creatives increases. To this end, this paper proposes a two-stage framework named Automated Creatives Quota (ACQ) to achieve the automatic creation and deactivation of ad creatives. ACQ dynamically allocates the creative quota across multiple advertisers to maximize the revenue of the ad platform. ACQ comprises two components: a prediction module to estimate the cost of a photo under different numbers of ad creatives, and an allocation module to decide the quota for photos considering their estimated costs in the prediction module. Specifically, in the prediction module, we develop a multi-task learning model based on an unbalanced binary tree to effectively mitigate the target variable imbalance problem. In the allocation module, we formulate the quota allocation problem as a multiple-choice knapsack problem (MCKP) and develop an efficient solver to solve such large-scale problems involving tens of millions of ads. We performed extensive offline and online experiments to validate the superiority of our proposed framework, which increased cost by 9.34%.
☆ Conservative Contextual Bandits: Beyond Linear Representations
Conservative Contextual Bandits (CCBs) address safety in sequential decision making by requiring that an agent's policy, along with minimizing regret, also satisfies a safety constraint: the performance is not worse than a baseline policy (e.g., the policy that the company has in production) by more than $(1+\alpha)$ factor. Prior work developed UCB-style algorithms in the multi-armed [Wu et al., 2016] and contextual linear [Kazerouni et al., 2017] settings. However, in practice the cost of the arms is often a non-linear function, and therefore existing UCB algorithms are ineffective in such settings. In this paper, we consider CCBs beyond the linear case and develop two algorithms $\mathtt{C-SquareCB}$ and $\mathtt{C-FastCB}$, using Inverse Gap Weighting (IGW) based exploration and an online regression oracle. We show that the safety constraint is satisfied with high probability and that the regret of $\mathtt{C-SquareCB}$ is sub-linear in horizon $T$, while the regret of $\mathtt{C-FastCB}$ is first-order and is sub-linear in $L^*$, the cumulative loss of the optimal policy. Subsequently, we use a neural network for function approximation and online gradient descent as the regression oracle to provide $\tilde{O}(\sqrt{KT} + K/\alpha) $ and $\tilde{O}(\sqrt{KL^*} + K (1 + 1/\alpha))$ regret bounds, respectively. Finally, we demonstrate the efficacy of our algorithms on real-world data and show that they significantly outperform the existing baseline while maintaining the performance guarantee.
☆ Query-Efficient Planning with Language Models
Planning in complex environments requires an agent to efficiently query a world model to find a feasible sequence of actions from start to goal. Recent work has shown that Large Language Models (LLMs), with their rich prior knowledge and reasoning capabilities, can potentially help with planning by searching over promising states and adapting to feedback from the world. In this paper, we propose and study two fundamentally competing frameworks that leverage LLMs for query-efficient planning. The first uses LLMs as a heuristic within a search-based planner to select promising nodes to expand and propose promising actions. The second uses LLMs as a generative planner to propose an entire sequence of actions from start to goal, query a world model, and adapt based on feedback. We show that while both approaches improve upon comparable baselines, using an LLM as a generative planner results in significantly fewer interactions. Our key finding is that the LLM as a planner can more rapidly adapt its planning strategies based on immediate feedback than LLM as a heuristic. We present evaluations and ablations on Robotouille and PDDL planning benchmarks and discuss connections to existing theory on query-efficient planning algorithms. Code is available at https://github.com/portal-cornell/llms-for-planning
comment: 11 pages (not including references or appendix); 13 figures (9 main paper, 4 appendix); (v1) preprint
☆ MoSH: Modeling Multi-Objective Tradeoffs with Soft and Hard Bounds
Countless science and engineering applications in multi-objective optimization (MOO) necessitate that decision-makers (DMs) select a Pareto-optimal solution which aligns with their preferences. Evaluating individual solutions is often expensive, necessitating cost-sensitive optimization techniques. Due to competing objectives, the space of trade-offs is also expansive -- thus, examining the full Pareto frontier may prove overwhelming to a DM. Such real-world settings generally have loosely-defined and context-specific desirable regions for each objective function that can aid in constraining the search over the Pareto frontier. We introduce a novel conceptual framework that operationalizes these priors using soft-hard functions, SHFs, which allow for the DM to intuitively impose soft and hard bounds on each objective -- which has been lacking in previous MOO frameworks. Leveraging a novel minimax formulation for Pareto frontier sampling, we propose a two-step process for obtaining a compact set of Pareto-optimal points which respect the user-defined soft and hard bounds: (1) densely sample the Pareto frontier using Bayesian optimization, and (2) sparsify the selected set to surface to the user, using robust submodular function optimization. We prove that (2) obtains the optimal compact Pareto-optimal set of points from (1). We further show that many practical problems fit within the SHF framework and provide extensive empirical validation on diverse domains, including brachytherapy, engineering design, and large language model personalization. Specifically, for brachytherapy, our approach returns a compact set of points with over 3% greater SHF-defined utility than the next best approach. Among the other diverse experiments, our approach consistently leads in utility, allowing the DM to reach >99% of their maximum possible desired utility within validation of 5 points.
☆ The Computational Limits of State-Space Models and Mamba via the Lens of Circuit Complexity
In this paper, we analyze the computational limitations of Mamba and State-space Models (SSMs) by using the circuit complexity framework. Despite Mamba's stateful design and recent attention as a strong candidate to outperform Transformers, we have demonstrated that both Mamba and SSMs with $\mathrm{poly}(n)$-precision and constant-depth layers reside within the $\mathsf{DLOGTIME}$-uniform $\mathsf{TC}^0$ complexity class. This result indicates Mamba has the same computational capabilities as Transformer theoretically, and it cannot solve problems like arithmetic formula problems, boolean formula value problems, and permutation composition problems if $\mathsf{TC}^0 \neq \mathsf{NC}^1$. Therefore, it challenges the assumption Mamba is more computationally expressive than Transformers. Our contributions include rigorous proofs showing that Selective SSM and Mamba architectures can be simulated by $\mathsf{DLOGTIME}$-uniform $\mathsf{TC}^0$ circuits, and they cannot solve problems outside $\mathsf{TC}^0$.
☆ Homogeneous Dynamics Space for Heterogeneous Humans
Analyses of human motion kinematics have achieved tremendous advances. However, the production mechanism, known as human dynamics, is still undercovered. In this paper, we aim to push data-driven human dynamics understanding forward. We identify a major obstacle to this as the heterogeneity of existing human motion understanding efforts. Specifically, heterogeneity exists in not only the diverse kinematics representations and hierarchical dynamics representations but also in the data from different domains, namely biomechanics and reinforcement learning. With an in-depth analysis of the existing heterogeneity, we propose to emphasize the beneath homogeneity: all of them represent the homogeneous fact of human motion, though from different perspectives. Given this, we propose Homogeneous Dynamics Space (HDyS) as a fundamental space for human dynamics by aggregating heterogeneous data and training a homogeneous latent space with inspiration from the inverse-forward dynamics procedure. Leveraging the heterogeneous representations and datasets, HDyS achieves decent mapping between human kinematics and dynamics. We demonstrate the feasibility of HDyS with extensive experiments and applications. The project page is https://foruck.github.io/HDyS.
comment: Cewu Lu and Yong-Lu Li are the corresponding authors
☆ Precise, Fast, and Low-cost Concept Erasure in Value Space: Orthogonal Complement Matters
The success of text-to-image generation enabled by diffuion models has imposed an urgent need to erase unwanted concepts, e.g., copyrighted, offensive, and unsafe ones, from the pre-trained models in a precise, timely, and low-cost manner. The twofold demand of concept erasure requires a precise removal of the target concept during generation (i.e., erasure efficacy), while a minimal impact on non-target content generation (i.e., prior preservation). Existing methods are either computationally costly or face challenges in maintaining an effective balance between erasure efficacy and prior preservation. To improve, we propose a precise, fast, and low-cost concept erasure method, called Adaptive Vaule Decomposer (AdaVD), which is training-free. This method is grounded in a classical linear algebraic orthogonal complement operation, implemented in the value space of each cross-attention layer within the UNet of diffusion models. An effective shift factor is designed to adaptively navigate the erasure strength, enhancing prior preservation without sacrificing erasure efficacy. Extensive experimental results show that the proposed AdaVD is effective at both single and multiple concept erasure, showing a 2- to 10-fold improvement in prior preservation as compared to the second best, meanwhile achieving the best or near best erasure efficacy, when comparing with both training-based and training-free state of the arts. AdaVD supports a series of diffusion models and downstream image generation tasks, the code is available on the project page: https://github.com/WYuan1001/AdaVD
☆ MMedPO: Aligning Medical Vision-Language Models with Clinical-Aware Multimodal Preference Optimization
The advancement of Large Vision-Language Models (LVLMs) has propelled their application in the medical field. However, Medical LVLMs (Med-LVLMs) encounter factuality challenges due to modality misalignment, where the models prioritize textual knowledge over visual input, leading to hallucinations that contradict information in medical images. Previous attempts to enhance modality alignment in Med-LVLMs through preference optimization have inadequately mitigated clinical relevance in preference data, making these samples easily distinguishable and reducing alignment effectiveness. To address this challenge, we propose MMedPO, a novel multimodal medical preference optimization approach that considers the clinical relevance of preference samples to enhance Med-LVLM alignment. MMedPO curates multimodal preference data by introducing two types of dispreference: (1) plausible hallucinations injected through target Med-LVLMs or GPT-4o to produce medically inaccurate responses, and (2) lesion region neglect achieved through local lesion-noising, disrupting visual understanding of critical areas. We then calculate clinical relevance for each sample based on scores from multiple Med-LLMs and visual tools, and integrate these scores into the preference optimization process as weights, enabling effective alignment. Our experiments demonstrate that MMedPO significantly enhances factual accuracy in Med-LVLMs, achieving substantial improvements over existing preference optimization methods by averaging 14.2% and 51.7% across the Med-VQA and report generation tasks. Our code are available in https://github.com/aiming-lab/MMedPO.
☆ Privacy-Preserving Large Language Models: Mechanisms, Applications, and Future Directions
The rapid advancement of large language models (LLMs) has revolutionized natural language processing, enabling applications in diverse domains such as healthcare, finance and education. However, the growing reliance on extensive data for training and inference has raised significant privacy concerns, ranging from data leakage to adversarial attacks. This survey comprehensively explores the landscape of privacy-preserving mechanisms tailored for LLMs, including differential privacy, federated learning, cryptographic protocols, and trusted execution environments. We examine their efficacy in addressing key privacy challenges, such as membership inference and model inversion attacks, while balancing trade-offs between privacy and model utility. Furthermore, we analyze privacy-preserving applications of LLMs in privacy-sensitive domains, highlighting successful implementations and inherent limitations. Finally, this survey identifies emerging research directions, emphasizing the need for novel frameworks that integrate privacy by design into the lifecycle of LLMs. By synthesizing state-of-the-art approaches and future trends, this paper provides a foundation for developing robust, privacy-preserving large language models that safeguard sensitive information without compromising performance.
♻ ☆ From Concept to Manufacturing: Evaluating Vision-Language Models for Engineering Design
Engineering design is undergoing a transformative shift with the advent of AI, marking a new era in how we approach product, system, and service planning. Large language models have demonstrated impressive capabilities in enabling this shift. Yet, with text as their only input modality, they cannot leverage the large body of visual artifacts that engineers have used for centuries and are accustomed to. This gap is addressed with the release of multimodal vision-language models (VLMs), such as GPT-4V, enabling AI to impact many more types of tasks. Our work presents a comprehensive evaluation of VLMs across a spectrum of engineering design tasks, categorized into four main areas: Conceptual Design, System-Level and Detailed Design, Manufacturing and Inspection, and Engineering Education Tasks. Specifically in this paper, we assess the capabilities of two VLMs, GPT-4V and LLaVA 1.6 34B, in design tasks such as sketch similarity analysis, CAD generation, topology optimization, manufacturability assessment, and engineering textbook problems. Through this structured evaluation, we not only explore VLMs' proficiency in handling complex design challenges but also identify their limitations in complex engineering design applications. Our research establishes a foundation for future assessments of vision language models. It also contributes a set of benchmark testing datasets, with more than 1000 queries, for ongoing advancements and applications in this field.
♻ ☆ Predictive Models in Sequential Recommendations: Bridging Performance Laws with Data Quality Insights
Sequential Recommendation (SR) plays a critical role in predicting users' sequential preferences. Despite its growing prominence in various industries, the increasing scale of SR models incurs substantial computational costs and unpredictability, challenging developers to manage resources efficiently. Under this predicament, Scaling Laws have achieved significant success by examining the loss as models scale up. However, there remains a disparity between loss and model performance, which is of greater concern in practical applications. Moreover, as data continues to expand, it incorporates repetitive and inefficient data. In response, we introduce the Performance Law for SR models, which aims to theoretically investigate and model the relationship between model performance and data quality. Specifically, we first fit the HR and NDCG metrics to transformer-based SR models. Subsequently, we propose Approximate Entropy (ApEn) to assess data quality, presenting a more nuanced approach compared to traditional data quantity metrics. Our method enables accurate predictions across various dataset scales and model sizes, demonstrating a strong correlation in large SR models and offering insights into achieving optimal performance for any given model configuration.
comment: 12 pages, 5 figures
♻ ☆ ZS4C: Zero-Shot Synthesis of Compilable Code for Incomplete Code Snippets using LLMs
Technical Q&A sites are valuable for software developers seeking knowledge, but the code snippets they provide are often uncompilable and incomplete due to unresolved types and missing libraries. This poses a challenge for users who wish to reuse or analyze these snippets. Existing methods either do not focus on creating compilable code or have low success rates. To address this, we propose ZS4C, a lightweight approach for zero-shot synthesis of compilable code from incomplete snippets using Large Language Models (LLMs). ZS4C operates in two stages: first, it uses an LLM, like GPT-3.5, to identify missing import statements in a snippet; second, it collaborates with a validator (e.g., compiler) to fix compilation errors caused by incorrect imports and syntax issues. We evaluated ZS4C on the StatType-SO benchmark and a new dataset, Python-SO, which includes 539 Python snippets from Stack Overflow across the 20 most popular Python libraries. ZS4C significantly outperforms existing methods, improving the compilation rate from 63% to 95.1% compared to the state-of-the-art SnR, marking a 50.1% improvement. On average, ZS4C can infer more accurate import statements (with an F1 score of 0.98) than SnR, with an improvement of 8.5% in the F1.
comment: This paper has been accepted and published in ACM Transactions on Software Engineering and Methodology (TOSEM), [2024], [https://dl.acm.org/doi/10.1145/3702979]
♻ ☆ Croissant: A Metadata Format for ML-Ready Datasets NeurIPS 2024
Data is a critical resource for machine learning (ML), yet working with data remains a key friction point. This paper introduces Croissant, a metadata format for datasets that creates a shared representation across ML tools, frameworks, and platforms. Croissant makes datasets more discoverable, portable, and interoperable, thereby addressing significant challenges in ML data management. Croissant is already supported by several popular dataset repositories, spanning hundreds of thousands of datasets, enabling easy loading into the most commonly-used ML frameworks, regardless of where the data is stored. Our initial evaluation by human raters shows that Croissant metadata is readable, understandable, complete, yet concise.
comment: Published at the NeurIPS 2024 Datasets and Benchmark Track. A shorter version appeared earlier in Proceedings of ACM SIGMOD/PODS'24 Data Management for End-to-End Machine Learning (DEEM) Workshop https://dl.acm.org/doi/10.1145/3650203.3663326
♻ ☆ Deep-Learning Based Docking Methods: Fair Comparisons to Conventional Docking Workflows
The diffusion learning method, DiffDock, for docking small-molecule ligands into protein binding sites was recently introduced. Results included comparisons to more conventional docking approaches, with DiffDock showing superior performance. Here, we employ a fully automatic workflow using the Surflex-Dock methods to generate a fair baseline for conventional docking approaches. Results were generated for the common and expected situation where a binding site location is known and also for the condition of an unknown binding site. For the known binding site condition, Surflex-Dock success rates at 2.0 Angstroms RMSD far exceeded those for DiffDock (Top-1/Top-5 success rates, respectively, were 68/81% compared with 45/51%). Glide performed with similar success rates (67/73%) to Surflex-Dock for the known binding site condition, and results for AutoDock Vina and Gnina followed this pattern. For the unknown binding site condition, using an automated method to identify multiple binding pockets, Surflex-Dock success rates again exceeded those of DiffDock, but by a somewhat lesser margin. DiffDock made use of roughly 17,000 co-crystal structures for learning (98% of PDBBind version 2020, pre-2019 structures) for a training set in order to predict on 363 test cases (2% of PDBBind 2020) from 2019 forward. DiffDock's performance was inextricably linked with the presence of near-neighbor cases of close to identical protein-ligand complexes in the training set for over half of the test set cases. DiffDock exhibited a 40 percentage point difference on near-neighbor cases (two-thirds of all test cases) compared with cases with no near-neighbor training case. DiffDock has apparently encoded a type of table-lookup during its learning process, rendering meaningful applications beyond its reach. Further, it does not perform even close to competitively with a competently run modern docking workflow.
comment: Post-Conclusion addendum added with additional reference and context, 19 pages including references and appendices, 7 figures
♻ ☆ LLM Pruning and Distillation in Practice: The Minitron Approach
We present a comprehensive report on compressing the Llama 3.1 8B and Mistral NeMo 12B models to 4B and 8B parameters, respectively, using pruning and distillation. We explore two distinct pruning strategies: (1) depth pruning and (2) joint hidden/attention/MLP (width) pruning, and evaluate the results on common benchmarks from the LM Evaluation Harness. The models are then aligned with NeMo Aligner and tested in instruct-tuned versions. This approach produces a compelling 4B model from Llama 3.1 8B and a state-of-the-art Mistral-NeMo-Minitron-8B (MN-Minitron-8B for brevity) model from Mistral NeMo 12B. We found that with no access to the original data, it is beneficial to slightly fine-tune teacher models on the distillation dataset. We open-source our base model weights on Hugging Face with a permissive license.
comment: v4: Update author order
♻ ☆ OW-VISCapTor: Abstractors for Open-World Video Instance Segmentation and Captioning SC
We propose the new task 'open-world video instance segmentation and captioning'. It requires to detect, segment, track and describe with rich captions never before seen objects. This challenging task can be addressed by developing "abstractors" which connect a vision model and a language foundation model. Concretely, we connect a multi-scale visual feature extractor and a large language model (LLM) by developing an object abstractor and an object-to-text abstractor. The object abstractor, consisting of a prompt encoder and transformer blocks, introduces spatially-diverse open-world object queries to discover never before seen objects in videos. An inter-query contrastive loss further encourages the diversity of object queries. The object-to-text abstractor is augmented with masked cross-attention and acts as a bridge between the object queries and a frozen LLM to generate rich and descriptive object-centric captions for each detected object. Our generalized approach surpasses the baseline that jointly addresses the tasks of open-world video instance segmentation and dense video object captioning by 13% on never before seen objects, and by 10% on object-centric captions.
comment: Project page: https://anwesachoudhuri.github.io/OpenWorldVISCap/
♻ ☆ Multi-Agent Quantum Reinforcement Learning using Evolutionary Optimization
Multi-Agent Reinforcement Learning is becoming increasingly more important in times of autonomous driving and other smart industrial applications. Simultaneously a promising new approach to Reinforcement Learning arises using the inherent properties of quantum mechanics, reducing the trainable parameters of a model significantly. However, gradient-based Multi-Agent Quantum Reinforcement Learning methods often have to struggle with barren plateaus, holding them back from matching the performance of classical approaches. We build upon an existing approach for gradient free Quantum Reinforcement Learning and propose three genetic variations with Variational Quantum Circuits for Multi-Agent Reinforcement Learning using evolutionary optimization. We evaluate our genetic variations in the Coin Game environment and also compare them to classical approaches. We showed that our Variational Quantum Circuit approaches perform significantly better compared to a neural network with a similar amount of trainable parameters. Compared to the larger neural network, our approaches archive similar results using $97.88\%$ less parameters.
♻ ☆ Wake Vision: A Tailored Dataset and Benchmark Suite for TinyML Computer Vision Applications
Tiny machine learning (TinyML) for low-power devices lacks robust datasets for development. We present Wake Vision, a large-scale dataset for person detection that contains over 6 million quality-filtered images. We provide two variants: Wake Vision (Large) and Wake Vision (Quality), leveraging the large variant for pretraining and knowledge distillation, while the higher-quality labels drive final model performance. The manually labeled validation and test sets reduce error rates from 7.8% to 2.2% compared to previous standards. In addition, we introduce five detailed benchmark sets to evaluate model performance in real-world scenarios, including varying lighting, camera distances, and demographic characteristics. Training with Wake Vision improves accuracy by 1.93% over existing datasets, demonstrating the importance of dataset quality for low-capacity models and dataset size for high-capacity models. The dataset, benchmarks, code, and models are available under the CC-BY 4.0 license, maintained by the Edge AI Foundation.
♻ ☆ GeoSAM: Fine-tuning SAM with Multi-Modal Prompts for Mobility Infrastructure Segmentation
In geographical image segmentation, performance is often constrained by the limited availability of training data and a lack of generalizability, particularly for segmenting mobility infrastructure such as roads, sidewalks, and crosswalks. Vision foundation models like the Segment Anything Model (SAM), pre-trained on millions of natural images, have demonstrated impressive zero-shot segmentation performance, providing a potential solution. However, SAM struggles with geographical images, such as aerial and satellite imagery, due to its training being confined to natural images and the narrow features and textures of these objects blending into their surroundings. To address these challenges, we propose Geographical SAM (GeoSAM), a SAM-based framework that fine-tunes SAM with automatically generated multi-modal prompts, combining point prompts from a pre-trained task-specific model as primary visual guidance and text prompts from a large language model as secondary semantic guidance to enhance model comprehension. GeoSAM outperforms existing approaches for mobility infrastructure segmentation in both familiar and completely unseen regions by at least 5\% in mIoU, representing a significant leap in leveraging foundation models to segment mobility infrastructure, including both road and pedestrian infrastructure in geographical images. The source code can be found in this GitHub Repository: https://github.com/rafiibnsultan/GeoSAM.
♻ ☆ FullStack Bench: Evaluating LLMs as Full Stack Coders
As the capabilities of code large language models (LLMs) continue to expand, their applications across diverse code intelligence domains are rapidly increasing. However, most existing datasets only evaluate limited application domains. To address this gap, we have developed a comprehensive code evaluation dataset FullStack Bench focusing on full-stack programming, which encompasses a wide range of application domains (e.g., basic programming, data analysis, software engineering, mathematics, and machine learning). Besides, to assess multilingual programming capabilities, in FullStack Bench, we design real-world instructions and corresponding unit test cases from 16 widely-used programming languages to reflect real-world usage scenarios rather than simple translations. Moreover, we also release an effective code sandbox execution tool (i.e., SandboxFusion) supporting various programming languages and packages to evaluate the performance of our FullStack Bench efficiently. Comprehensive experimental results on our FullStack Bench demonstrate the necessity and effectiveness of our FullStack Bench and SandboxFusion.
comment: 26 pages
♻ ☆ Can tweets predict article retractions? A comparison between human and LLM labelling
Quickly detecting problematic research articles is crucial to safeguarding the integrity of scientific research. This study explores whether Twitter mentions of retracted articles can signal potential problems with the articles prior to their retraction, potentially serving as an early warning system for scholars. To investigate this, we analysed a dataset of 4,354 Twitter mentions associated with 504 retracted articles. The effectiveness of Twitter mentions in predicting article retractions was evaluated by both manual and Large Language Model (LLM) labelling. Manual labelling results indicated that 25.7% of tweets signalled problems before retraction. Using the manual labelling results as the baseline, we found that LLMs (GPT-4o-mini, Gemini 1.5 Flash, and Claude-3.5-Haiku) outperformed lexicon-based sentiment analysis tools (e.g., TextBlob) in detecting potential problems, suggesting that automatic detection of problematic articles from social media using LLMs is technically feasible. Nevertheless, since only a small proportion of retracted articles (11.1%) were criticised on Twitter prior to retraction, such automatic systems would detect only a minority of problematic articles. Overall, this study offers insights into how social media data, coupled with emerging generative AI techniques, can support research integrity.
comment: 19 pages
♻ ☆ How Transformers Solve Propositional Logic Problems: A Mechanistic Analysis
Large language models (LLMs) have shown amazing performance on tasks that require planning and reasoning. Motivated by this, we investigate the internal mechanisms that underpin a network's ability to perform complex logical reasoning. We first construct a synthetic propositional logic problem that serves as a concrete test-bed for network training and evaluation. Crucially, this problem demands nontrivial planning to solve. We perform our study on two fronts. First, we pursue an understanding of precisely how a three-layer transformer, trained from scratch and attains perfect test accuracy, solves this problem. We are able to identify certain "planning" and "reasoning" mechanisms in the network that necessitate cooperation between the attention blocks to implement the desired logic. Second, we study how pretrained LLMs, namely Mistral-7B and Gemma-2-9B, solve this problem. We characterize their reasoning circuits through causal intervention experiments, providing necessity and sufficiency evidence for the circuits. We find evidence suggesting that the two models' latent reasoning strategies are surprisingly similar, and human-like. Overall, our work systemically uncovers novel aspects of small and large transformers, and continues the study of how they plan and reason.
♻ ☆ Code-as-Monitor: Constraint-aware Visual Programming for Reactive and Proactive Robotic Failure Detection
Automatic detection and prevention of open-set failures are crucial in closed-loop robotic systems. Recent studies often struggle to simultaneously identify unexpected failures reactively after they occur and prevent foreseeable ones proactively. To this end, we propose Code-as-Monitor (CaM), a novel paradigm leveraging the vision-language model (VLM) for both open-set reactive and proactive failure detection. The core of our method is to formulate both tasks as a unified set of spatio-temporal constraint satisfaction problems and use VLM-generated code to evaluate them for real-time monitoring. To enhance the accuracy and efficiency of monitoring, we further introduce constraint elements that abstract constraint-related entities or their parts into compact geometric elements. This approach offers greater generality, simplifies tracking, and facilitates constraint-aware visual programming by leveraging these elements as visual prompts. Experiments show that CaM achieves a 28.7% higher success rate and reduces execution time by 31.8% under severe disturbances compared to baselines across three simulators and a real-world setting. Moreover, CaM can be integrated with open-loop control policies to form closed-loop systems, enabling long-horizon tasks in cluttered scenes with dynamic environments.
comment: Project page: https://zhoues.github.io/Code-as-Monitor/
♻ ☆ Social Media Informatics for Sustainable Cities and Societies: An Overview of the Applications, associated Challenges, and Potential Solutions
In the modern world, our cities and societies face several technological and societal challenges, such as rapid urbanization, global warming & climate change, the digital divide, and social inequalities, increasing the need for more sustainable cities and societies. Addressing these challenges requires a multifaceted approach involving all the stakeholders, sustainable planning, efficient resource management, innovative solutions, and modern technologies. Like other modern technologies, social media informatics also plays its part in developing more sustainable and resilient cities and societies. Despite its limitations, social media informatics has proven very effective in various sustainable cities and society applications. In this paper, we review and analyze the role of social media informatics in sustainable cities and society by providing a detailed overview of its applications, associated challenges, and potential solutions. This work is expected to provide a baseline for future research in the domain.
comment: 35 pages, 3 tables, and 4 figures
♻ ☆ APOLLO: SGD-like Memory, AdamW-level Performance
Large language models (LLMs) are notoriously memory-intensive during training, particularly with the popular AdamW optimizer. This memory burden necessitates using more or higher-end GPUs or reducing batch sizes, limiting training scalability and throughput. To address this, various memory-efficient optimizers have been proposed to reduce optimizer memory usage. However, they face critical challenges: (i) reliance on costly SVD operations; (ii) significant performance trade-offs compared to AdamW; and (iii) still substantial optimizer memory overhead to maintain competitive performance. In this work, we identify that AdamW's learning rate adaptation rule can be effectively coarsened as a structured learning rate update. Based on this insight, we propose Approximated Gradient Scaling for Memory-Efficient LLM Optimization (APOLLO), which approximates learning rate scaling using an auxiliary low-rank optimizer state based on pure random projection. This structured learning rate update rule makes APOLLO highly tolerant to further memory reductions while delivering comparable pre-training performance. Even its rank-1 variant, APOLLO-Mini, achieves superior pre-training performance compared to AdamW with SGD-level memory costs. Extensive experiments demonstrate that the APOLLO series performs on-par with or better than AdamW, while achieving greater memory savings by nearly eliminating the optimization states of AdamW. These savings provide significant system-level benefits: (1) Enhanced Throughput: 3x throughput on an 8xA100-80GB setup compared to AdamW by supporting 4x larger batch sizes. (2) Improved Model Scalability: Pre-training LLaMA-13B with naive DDP on A100-80GB GPUs without system-level optimizations. (3) Low-End GPU Friendly Pre-training: Pre-training LLaMA-7B on a single GPU using less than 12 GB of memory with weight quantization.
comment: Preprint
♻ ☆ Enhancing predictive imaging biomarker discovery through treatment effect analysis WACV 2025
Identifying predictive covariates, which forecast individual treatment effectiveness, is crucial for decision-making across different disciplines such as personalized medicine. These covariates, referred to as biomarkers, are extracted from pre-treatment data, often within randomized controlled trials, and should be distinguished from prognostic biomarkers, which are independent of treatment assignment. Our study focuses on discovering predictive imaging biomarkers, specific image features, by leveraging pre-treatment images to uncover new causal relationships. Unlike labor-intensive approaches relying on handcrafted features prone to bias, we present a novel task of directly learning predictive features from images. We propose an evaluation protocol to assess a model's ability to identify predictive imaging biomarkers and differentiate them from purely prognostic ones by employing statistical testing and a comprehensive analysis of image feature attribution. We explore the suitability of deep learning models originally developed for estimating the conditional average treatment effect (CATE) for this task, which have been assessed primarily for their precision of CATE estimation while overlooking the evaluation of imaging biomarker discovery. Our proof-of-concept analysis demonstrates the feasibility and potential of our approach in discovering and validating predictive imaging biomarkers from synthetic outcomes and real-world image datasets. Our code is available at \url{https://github.com/MIC-DKFZ/predictive_image_biomarker_analysis}.
comment: Accepted to WACV 2025
♻ ☆ Synergistic Development of Perovskite Memristors and Algorithms for Robust Analog Computing
Analog computing using non-volatile memristors has emerged as a promising solution for energy-efficient deep learning. New materials, like perovskites-based memristors are recently attractive due to their cost-effectiveness, energy efficiency and flexibility. Yet, challenges in material diversity and immature fabrications require extensive experimentation for device development. Moreover, significant non-idealities in these memristors often impede them for computing. Here, we propose a synergistic methodology to concurrently optimize perovskite memristor fabrication and develop robust analog DNNs that effectively address the inherent non-idealities of these memristors. Employing Bayesian optimization (BO) with a focus on usability, we efficiently identify optimal materials and fabrication conditions for perovskite memristors. Meanwhile, we developed "BayesMulti", a DNN training strategy utilizing BO-guided noise injection to improve the resistance of analog DNNs to memristor imperfections. Our approach theoretically ensures that within a certain range of parameter perturbations due to memristor non-idealities, the prediction outcomes remain consistent. Our integrated approach enables use of analog computing in much deeper and wider networks, which significantly outperforms existing methods in diverse tasks like image classification, autonomous driving, species identification, and large vision-language models, achieving up to 100-fold improvements. We further validate our methodology on a 10$\times$10 optimized perovskite memristor crossbar, demonstrating high accuracy in a classification task and low energy consumption. This study offers a versatile solution for efficient optimization of various analog computing systems, encompassing both devices and algorithms.
♻ ☆ CIC: A Framework for Culturally-Aware Image Captioning IJCAI 2024
Image Captioning generates descriptive sentences from images using Vision-Language Pre-trained models (VLPs) such as BLIP, which has improved greatly. However, current methods lack the generation of detailed descriptive captions for the cultural elements depicted in the images, such as the traditional clothing worn by people from Asian cultural groups. In this paper, we propose a new framework, Culturally-aware Image Captioning (CIC), that generates captions and describes cultural elements extracted from cultural visual elements in images representing cultures. Inspired by methods combining visual modality and Large Language Models (LLMs) through appropriate prompts, our framework (1) generates questions based on cultural categories from images, (2) extracts cultural visual elements from Visual Question Answering (VQA) using generated questions, and (3) generates culturally-aware captions using LLMs with the prompts. Our human evaluation conducted on 45 participants from 4 different cultural groups with a high understanding of the corresponding culture shows that our proposed framework generates more culturally descriptive captions when compared to the image captioning baseline based on VLPs. Resources can be found at https://shane3606.github.io/cic..
comment: Accepted in IJCAI 2024
♻ ☆ AI Cyber Risk Benchmark: Automated Exploitation Capabilities
We introduce a new benchmark for assessing AI models' capabilities and risks in automated software exploitation, focusing on their ability to detect and exploit vulnerabilities in real-world software systems. Using DARPA's AI Cyber Challenge (AIxCC) framework and the Nginx challenge project, a deliberately modified version of the widely used Nginx web server, we evaluate several leading language models, including OpenAI's o1-preview and o1-mini, Anthropic's Claude-3.5-sonnet-20241022 and Claude-3.5-sonnet-20240620, Google DeepMind's Gemini-1.5-pro, and OpenAI's earlier GPT-4o model. Our findings reveal that these models vary significantly in their success rates and efficiency, with o1-preview achieving the highest success rate of 64.71 percent and o1-mini and Claude-3.5-sonnet-20241022 providing cost-effective but less successful alternatives. This benchmark establishes a foundation for systematically evaluating the AI cyber risk posed by automated exploitation tools.
♻ ☆ Burning RED: Unlocking Subtask-Driven Reinforcement Learning and Risk-Awareness in Average-Reward Markov Decision Processes
Average-reward Markov decision processes (MDPs) provide a foundational framework for sequential decision-making under uncertainty. However, average-reward MDPs have remained largely unexplored in reinforcement learning (RL) settings, with the majority of RL-based efforts having been allocated to episodic and discounted MDPs. In this work, we study a unique structural property of average-reward MDPs and utilize it to introduce Reward-Extended Differential (or RED) reinforcement learning: a novel RL framework that can be used to effectively and efficiently solve various subtasks simultaneously in the average-reward setting. We introduce a family of RED learning algorithms for prediction and control, including proven-convergent algorithms for the tabular case. We then showcase the power of these algorithms by demonstrating how they can be used to learn a policy that optimizes, for the first time, the well-known conditional value-at-risk (CVaR) risk measure in a fully-online manner, without the use of an explicit bi-level optimization scheme or an augmented state-space.
♻ ☆ Improving the Efficiency of Visually Augmented Language Models COLING 2025
Despite the impressive performance of autoregressive Language Models (LM) it has been shown that due to reporting bias, LMs lack visual knowledge, i.e. they do not know much about the visual world and its properties. To augment LMs with visual knowledge, existing solutions often rely on explicit images, requiring time-consuming retrieval or image generation systems. This paper shows that explicit images are not necessary to visually augment an LM. Instead, we use visually-grounded text representations obtained from the well-known CLIP multimodal system. For a fair comparison, we modify VALM, a visually-augmented LM which uses image retrieval and representation, to work directly with visually-grounded text representations. We name this new model BLIND-VALM. We show that BLIND-VALM performs on par with VALM for Visual Language Understanding (VLU), Natural Language Understanding (NLU) and Language Modeling tasks, despite being significantly more efficient and simpler. We also show that scaling up our model within the compute budget of VALM, either increasing the model or pre-training corpus size, we outperform VALM for all the evaluation tasks.
comment: COLING 2025
♻ ☆ A Cognac shot to forget bad memories: Corrective Unlearning in GNNs
Graph Neural Networks (GNNs) are increasingly being used for a variety of ML applications on graph data. Because graph data does not follow the independently and identically distributed (i.i.d.) assumption, adversarial manipulations or incorrect data can propagate to other data points through message passing, which deteriorates the model's performance. To allow model developers to remove the adverse effects of manipulated entities from a trained GNN, we study the recently formulated problem of Corrective Unlearning. We find that current graph unlearning methods fail to unlearn the effect of manipulations even when the whole manipulated set is known. We introduce a new graph unlearning method, Cognac, which can unlearn the effect of the manipulation set even when only 5% of it is identified. It recovers most of the performance of a strong oracle with fully corrected training data, even beating retraining from scratch without the deletion set while being 8x more efficient. We hope our work assists GNN developers in mitigating harmful effects caused by issues in real-world data post-training. Our code is publicly available at https://github.com/varshitakolipaka/corrective-unlearning-for-gnns
♻ ☆ Similarity Equivariant Graph Neural Networks for Homogenization of Metamaterials
Soft, porous mechanical metamaterials exhibit pattern transformations that may have important applications in soft robotics, sound reduction and biomedicine. To design these innovative materials, it is important to be able to simulate them accurately and quickly, in order to tune their mechanical properties. Since conventional simulations using the finite element method entail a high computational cost, in this article we aim to develop a machine learning-based approach that scales favorably to serve as a surrogate model. To ensure that the model is also able to handle various microstructures, including those not encountered during training, we include the microstructure as part of the network input. Therefore, we introduce a graph neural network that predicts global quantities (energy, stress stiffness) as well as the pattern transformations that occur (the kinematics). To make our model as accurate and data-efficient as possible, various symmetries are incorporated into the model. The starting point is an E(n)-equivariant graph neural network (which respects translation, rotation and reflection) that has periodic boundary conditions (i.e., it is in-/equivariant with respect to the choice of RVE), is scale in-/equivariant, can simulate large deformations, and can predict scalars, vectors as well as second and fourth order tensors (specifically energy, stress and stiffness). The incorporation of scale equivariance makes the model equivariant with respect to the similarities group, of which the Euclidean group E(n) is a subgroup. We show that this network is more accurate and data-efficient than graph neural networks with fewer symmetries. To create an efficient graph representation of the finite element discretization, we use only the internal geometrical hole boundaries from the finite element mesh to achieve a better speed-up and scaling with the mesh size.
comment: 60 pages, 22 figures. Submitted to CMAME (Computer Methods in Applied Mechanics and Engineering)
♻ ☆ From Pixels to Words: Leveraging Explainability in Face Recognition through Interactive Natural Language Processing
Face Recognition (FR) has advanced significantly with the development of deep learning, achieving high accuracy in several applications. However, the lack of interpretability of these systems raises concerns about their accountability, fairness, and reliability. In the present study, we propose an interactive framework to enhance the explainability of FR models by combining model-agnostic Explainable Artificial Intelligence (XAI) and Natural Language Processing (NLP) techniques. The proposed framework is able to accurately answer various questions of the user through an interactive chatbot. In particular, the explanations generated by our proposed method are in the form of natural language text and visual representations, which for example can describe how different facial regions contribute to the similarity measure between two faces. This is achieved through the automatic analysis of the output's saliency heatmaps of the face images and a BERT question-answering model, providing users with an interface that facilitates a comprehensive understanding of the FR decisions. The proposed approach is interactive, allowing the users to ask questions to get more precise information based on the user's background knowledge. More importantly, in contrast to previous studies, our solution does not decrease the face recognition performance. We demonstrate the effectiveness of the method through different experiments, highlighting its potential to make FR systems more interpretable and user-friendly, especially in sensitive applications where decision-making transparency is crucial.
♻ ☆ Large Language Models and Games: A Survey and Roadmap
Recent years have seen an explosive increase in research on large language models (LLMs), and accompanying public engagement on the topic. While starting as a niche area within natural language processing, LLMs have shown remarkable potential across a broad range of applications and domains, including games. This paper surveys the current state of the art across the various applications of LLMs in and for games, and identifies the different roles LLMs can take within a game. Importantly, we discuss underexplored areas and promising directions for future uses of LLMs in games and we reconcile the potential and limitations of LLMs within the games domain. As the first comprehensive survey and roadmap at the intersection of LLMs and games, we are hopeful that this paper will serve as the basis for groundbreaking research and innovation in this exciting new field.
comment: Accepted for publication at the IEEE Transactions on Games (19 pages, 6 figures)
♻ ☆ Words2Contact: Identifying Support Contacts from Verbal Instructions Using Foundation Models
This paper presents Words2Contact, a language-guided multi-contact placement pipeline leveraging large language models and vision language models. Our method is a key component for language-assisted teleoperation and human-robot cooperation, where human operators can instruct the robots where to place their support contacts before whole-body reaching or manipulation using natural language. Words2Contact transforms the verbal instructions of a human operator into contact placement predictions; it also deals with iterative corrections, until the human is satisfied with the contact location identified in the robot's field of view. We benchmark state-of-the-art LLMs and VLMs for size and performance in contact prediction. We demonstrate the effectiveness of the iterative correction process, showing that users, even naive, quickly learn how to instruct the system to obtain accurate locations. Finally, we validate Words2Contact in real-world experiments with the Talos humanoid robot, instructed by human operators to place support contacts on different locations and surfaces to avoid falling when reaching for distant objects.
♻ ☆ A Survey on Multilingual Large Language Models: Corpora, Alignment, and Bias
Based on the foundation of Large Language Models (LLMs), Multilingual LLMs (MLLMs) have been developed to address the challenges faced in multilingual natural language processing, hoping to achieve knowledge transfer from high-resource languages to low-resource languages. However, significant limitations and challenges still exist, such as language imbalance, multilingual alignment, and inherent bias. In this paper, we aim to provide a comprehensive analysis of MLLMs, delving deeply into discussions surrounding these critical issues. First of all, we start by presenting an overview of MLLMs, covering their evolutions, key techniques, and multilingual capacities. Secondly, we explore the multilingual training corpora of MLLMs and the multilingual datasets oriented for downstream tasks that are crucial to enhance the cross-lingual capability of MLLMs. Thirdly, we survey the state-of-the-art studies of multilingual representations and investigate whether the current MLLMs can learn a universal language representation. Fourthly, we discuss bias on MLLMs, including its categories, evaluation metrics, and debiasing techniques. Finally, we discuss existing challenges and point out promising research directions of MLLMs.
comment: The article has been accepted by Frontiers of Computer Science (FCS), with the DOI: {10.1007/s11704-024-40579-4}
♻ ☆ Imitating Language via Scalable Inverse Reinforcement Learning NeurIPS 2024
The majority of language model training builds on imitation learning. It covers pretraining, supervised fine-tuning, and affects the starting conditions for reinforcement learning from human feedback (RLHF). The simplicity and scalability of maximum likelihood estimation (MLE) for next token prediction led to its role as predominant paradigm. However, the broader field of imitation learning can more effectively utilize the sequential structure underlying autoregressive generation. We focus on investigating the inverse reinforcement learning (IRL) perspective to imitation, extracting rewards and directly optimizing sequences instead of individual token likelihoods and evaluate its benefits for fine-tuning large language models. We provide a new angle, reformulating inverse soft-Q-learning as a temporal difference regularized extension of MLE. This creates a principled connection between MLE and IRL and allows trading off added complexity with increased performance and diversity of generations in the supervised fine-tuning (SFT) setting. We find clear advantages for IRL-based imitation, in particular for retaining diversity while maximizing task performance, rendering IRL a strong alternative on fixed SFT datasets even without online data generation. Our analysis of IRL-extracted reward functions further indicates benefits for more robust reward functions via tighter integration of supervised and preference-based LLM post-training.
comment: Published at NeurIPS 2024
♻ ☆ Self-supervised Monocular Depth and Pose Estimation for Endoscopy with Generative Latent Priors
Accurate 3D mapping in endoscopy enables quantitative, holistic lesion characterization within the gastrointestinal (GI) tract, requiring reliable depth and pose estimation. However, endoscopy systems are monocular, and existing methods relying on synthetic datasets or complex models often lack generalizability in challenging endoscopic conditions. We propose a robust self-supervised monocular depth and pose estimation framework that incorporates a Generative Latent Bank and a Variational Autoencoder (VAE). The Generative Latent Bank leverages extensive depth scenes from natural images to condition the depth network, enhancing realism and robustness of depth predictions through latent feature priors. For pose estimation, we reformulate it within a VAE framework, treating pose transitions as latent variables to regularize scale, stabilize z-axis prominence, and improve x-y sensitivity. This dual refinement pipeline enables accurate depth and pose predictions, effectively addressing the GI tract's complex textures and lighting. Extensive evaluations on SimCol and EndoSLAM datasets confirm our framework's superior performance over published self-supervised methods in endoscopic depth and pose estimation.
♻ ☆ Remove that Square Root: A New Efficient Scale-Invariant Version of AdaGrad
Adaptive methods are extremely popular in machine learning as they make learning rate tuning less expensive. This paper introduces a novel optimization algorithm named KATE, which presents a scale-invariant adaptation of the well-known AdaGrad algorithm. We prove the scale-invariance of KATE for the case of Generalized Linear Models. Moreover, for general smooth non-convex problems, we establish a convergence rate of $O \left(\frac{\log T}{\sqrt{T}} \right)$ for KATE, matching the best-known ones for AdaGrad and Adam. We also compare KATE to other state-of-the-art adaptive algorithms Adam and AdaGrad in numerical experiments with different problems, including complex machine learning tasks like image classification and text classification on real data. The results indicate that KATE consistently outperforms AdaGrad and matches/surpasses the performance of Adam in all considered scenarios.
comment: 31 pages, 12 figures
♻ ☆ LOLA -- An Open-Source Massively Multilingual Large Language Model
This paper presents LOLA, a massively multilingual large language model trained on more than 160 languages using a sparse Mixture-of-Experts Transformer architecture. Our architectural and implementation choices address the challenge of harnessing linguistic diversity while maintaining efficiency and avoiding the common pitfalls of multilinguality. Our analysis of the evaluation results shows competitive performance in natural language generation and understanding tasks. Additionally, we demonstrate how the learned expert-routing mechanism exploits implicit phylogenetic linguistic patterns to potentially alleviate the curse of multilinguality. We provide an in-depth look at the training process, an analysis of the datasets, and a balanced exploration of the model's strengths and limitations. As an open-source model, LOLA promotes reproducibility and serves as a robust foundation for future research. Our findings enable the development of compute-efficient multilingual models with strong, scalable performance across languages.
♻ ☆ Creativity in AI: Progresses and Challenges
Creativity is the ability to produce novel, useful, and surprising ideas, and has been widely studied as a crucial aspect of human cognition. Machine creativity on the other hand has been a long-standing challenge. With the rise of advanced generative AI, there has been renewed interest and debate regarding AI's creative capabilities. Therefore, it is imperative to revisit the state of creativity in AI and identify key progresses and remaining challenges. In this work, we survey leading works studying the creative capabilities of AI systems, focusing on creative problem-solving, linguistic, artistic, and scientific creativity. Our review suggests that while the latest AI models are largely capable of producing linguistically and artistically creative outputs such as poems, images, and musical pieces, they struggle with tasks that require creative problem-solving, abstract thinking and compositionality and their generations suffer from a lack of diversity, originality, long-range incoherence and hallucinations. We also discuss key questions concerning copyright and authorship issues with generative models. Furthermore, we highlight the need for a comprehensive evaluation of creativity that is process-driven and considers several dimensions of creativity. Finally, we propose future research directions to improve the creativity of AI outputs, drawing inspiration from cognitive science and psychology.
comment: minor updates to content + figures
♻ ☆ A philosophical and ontological perspective on Artificial General Intelligence and the Metaverse
This paper leverages various philosophical and ontological frameworks to explore the concept of embodied artificial general intelligence (AGI), its relationship to human consciousness, and the key role of the metaverse in facilitating this relationship. Several theoretical frameworks underpin this exploration, such as embodied cognition, Michael Levin's computational boundary of a "Self," Donald D. Hoffman's Interface Theory of Perception, and Bernardo Kastrup's analytical idealism, which lead to considering our perceived outer reality as a symbolic representation of alternate inner states of being, and where AGI could embody a different form of consciousness with a larger computational boundary. The paper further discusses the developmental stages of AGI, the requirements for the emergence of an embodied AGI, the importance of a calibrated symbolic interface for AGI, and the key role played by the metaverse, decentralized systems, open-source blockchain technology, as well as open-source AI research. It also explores the idea of a feedback loop between AGI and human users in metaverse spaces as a tool for AGI calibration, as well as the role of local homeostasis and decentralized governance as preconditions for achieving a stable embodied AGI. The paper concludes by emphasizing the importance of achieving a certain degree of harmony in human relations and recognizing the interconnectedness of humanity at a global level, as key prerequisites for the emergence of a stable embodied AGI.
comment: Presented at the conference second international conference on human-centred AI ethics: seeing the human in the artificial (HCAIE 2023): https://ethics-ai.eu/hcaie2023/
♻ ☆ System 2 Reasoning via Generality and Adaptation NeurIPS 2024
While significant progress has been made in task-specific applications, current models struggle with deep reasoning, generality, and adaptation -- key components of System 2 reasoning that are crucial for achieving Artificial General Intelligence (AGI). Despite the promise of approaches such as program synthesis, language models, and transformers, these methods often fail to generalize beyond their training data and to adapt to novel tasks, limiting their ability to perform human-like reasoning. This paper explores the limitations of existing approaches in achieving advanced System 2 reasoning and highlights the importance of generality and adaptation for AGI. Moreover, we propose four key research directions to address these gaps: (1) learning human intentions from action sequences, (2) combining symbolic and neural models, (3) meta-learning for unfamiliar environments, and (4) reinforcement learning to reason multi-step. Through these directions, we aim to advance the ability to generalize and adapt, bringing computational models closer to the reasoning capabilities required for AGI.
comment: Accepted by NeurIPS 2024 Workshop on System 2 Reasoning at Scale
♻ ☆ DNN Task Assignment in UAV Networks: A Generative AI Enhanced Multi-Agent Reinforcement Learning Approach
Unmanned Aerial Vehicles (UAVs) possess high mobility and flexible deployment capabilities, prompting the development of UAVs for various application scenarios within the Internet of Things (IoT). The unique capabilities of UAVs give rise to increasingly critical and complex tasks in uncertain and potentially harsh environments. The substantial amount of data generated from these applications necessitates processing and analysis through deep neural networks (DNNs). However, UAVs encounter challenges due to their limited computing resources when managing DNN models. This paper presents a joint approach that combines multiple-agent reinforcement learning (MARL) and generative diffusion models (GDM) for assigning DNN tasks to a UAV swarm, aimed at reducing latency from task capture to result output. To address these challenges, we first consider the task size of the target area to be inspected and the shortest flying path as optimization constraints, employing a greedy algorithm to resolve the subproblem with a focus on minimizing the UAV's flying path and the overall system cost. In the second stage, we introduce a novel DNN task assignment algorithm, termed GDM-MADDPG, which utilizes the reverse denoising process of GDM to replace the actor network in multi-agent deep deterministic policy gradient (MADDPG). This approach generates specific DNN task assignment actions based on agents' observations in a dynamic environment. Simulation results indicate that our algorithm performs favorably compared to benchmarks in terms of path planning, Age of Information (AoI), energy consumption, and task load balancing.
♻ ☆ PPFlow: Target-aware Peptide Design with Torsional Flow Matching
Therapeutic peptides have proven to have great pharmaceutical value and potential in recent decades. However, methods of AI-assisted peptide drug discovery are not fully explored. To fill the gap, we propose a target-aware peptide design method called \textsc{PPFlow}, based on conditional flow matching on torus manifolds, to model the internal geometries of torsion angles for the peptide structure design. Besides, we establish a protein-peptide binding dataset named PPBench2024 to fill the void of massive data for the task of structure-based peptide drug design and to allow the training of deep learning methods. Extensive experiments show that PPFlow reaches state-of-the-art performance in tasks of peptide drug generation and optimization in comparison with baseline models, and can be generalized to other tasks including docking and side-chain packing.
comment: 18 pages
♻ ☆ Dynamical similarity analysis uniquely captures how computations develop in RNNs
Methods for analyzing representations in neural systems have become a popular tool in both neuroscience and mechanistic interpretability. Having measures to compare how similar activations of neurons are across conditions, architectures, and species, gives us a scalable way of learning how information is transformed within different neural networks. In contrast to this trend, recent investigations have revealed how some metrics can respond to spurious signals and hence give misleading results. To identify the most reliable metric and understand how measures could be improved, it is going to be important to identify specific test cases which can serve as benchmarks. Here we propose that the phenomena of compositional learning in recurrent neural networks (RNNs) allows us to build a test case for dynamical representation alignment metrics. By implementing this case, we show it enables us to test whether metrics can identify representations which gradually develop throughout learning and probe whether representations identified by metrics are relevant to computations executed by networks. By building both an attractor- and RNN-based test case, we show that the new Dynamical Similarity Analysis (DSA) is more noise robust and identifies behaviorally relevant representations more reliably than prior metrics (Procrustes, CKA). We also show how test cases can be used beyond evaluating metrics to study new architectures. Specifically, results from applying DSA to modern (Mamba) state space models, suggest that, in contrast to RNNs, these models may not exhibit changes to their recurrent dynamics due to their expressiveness. Overall, by developing test cases, we show DSA's exceptional ability to detect compositional dynamical motifs, thereby enhancing our understanding of how computations unfold in RNNs.
comment: 19 pages, 11 figures
♻ ☆ Improved GUI Grounding via Iterative Narrowing
Graphical User Interface (GUI) grounding plays a crucial role in enhancing the capabilities of Vision-Language Model (VLM) agents. While general VLMs, such as GPT-4V, demonstrate strong performance across various tasks, their proficiency in GUI grounding remains suboptimal. Recent studies have focused on fine-tuning these models specifically for one-shot GUI grounding, yielding significant improvements over baseline performance. We introduce a visual prompting framework that employs an iterative narrowing mechanism to further improve the performance of both general and fine-tuned models in GUI grounding. For evaluation, we tested our method on a comprehensive benchmark comprising various UI platforms and provided the code to reproduce our results.
comment: Code available at https://github.com/ant-8/GUI-Grounding-via-Iterative-Narrowing
♻ ☆ Guiding a Diffusion Model with a Bad Version of Itself NeurIPS 2024
The primary axes of interest in image-generating diffusion models are image quality, the amount of variation in the results, and how well the results align with a given condition, e.g., a class label or a text prompt. The popular classifier-free guidance approach uses an unconditional model to guide a conditional model, leading to simultaneously better prompt alignment and higher-quality images at the cost of reduced variation. These effects seem inherently entangled, and thus hard to control. We make the surprising observation that it is possible to obtain disentangled control over image quality without compromising the amount of variation by guiding generation using a smaller, less-trained version of the model itself rather than an unconditional model. This leads to significant improvements in ImageNet generation, setting record FIDs of 1.01 for 64x64 and 1.25 for 512x512, using publicly available networks. Furthermore, the method is also applicable to unconditional diffusion models, drastically improving their quality.
comment: NeurIPS 2024
♻ ☆ Large Language Model Benchmarks in Medical Tasks
With the increasing application of large language models (LLMs) in the medical domain, evaluating these models' performance using benchmark datasets has become crucial. This paper presents a comprehensive survey of various benchmark datasets employed in medical LLM tasks. These datasets span multiple modalities including text, image, and multimodal benchmarks, focusing on different aspects of medical knowledge such as electronic health records (EHRs), doctor-patient dialogues, medical question-answering, and medical image captioning. The survey categorizes the datasets by modality, discussing their significance, data structure, and impact on the development of LLMs for clinical tasks such as diagnosis, report generation, and predictive decision support. Key benchmarks include MIMIC-III, MIMIC-IV, BioASQ, PubMedQA, and CheXpert, which have facilitated advancements in tasks like medical report generation, clinical summarization, and synthetic data generation. The paper summarizes the challenges and opportunities in leveraging these benchmarks for advancing multimodal medical intelligence, emphasizing the need for datasets with a greater degree of language diversity, structured omics data, and innovative approaches to synthesis. This work also provides a foundation for future research in the application of LLMs in medicine, contributing to the evolving field of medical artificial intelligence.
comment: 25 pages, 5 tables
♻ ☆ Assessing the potential of LLM-assisted annotation for corpus-based pragmatics and discourse analysis: The case of apology
Certain forms of linguistic annotation, like part of speech and semantic tagging, can be automated with high accuracy. However, manual annotation is still necessary for complex pragmatic and discursive features that lack a direct mapping to lexical forms. This manual process is time-consuming and error-prone, limiting the scalability of function-to-form approaches in corpus linguistics. To address this, our study explores the possibility of using large language models (LLMs) to automate pragma-discursive corpus annotation. We compare GPT-3.5 (the model behind the free-to-use version of ChatGPT), GPT-4 (the model underpinning the precise mode of Bing chatbot), and a human coder in annotating apology components in English based on the local grammar framework. We find that GPT-4 outperformed GPT-3.5, with accuracy approaching that of a human coder. These results suggest that LLMs can be successfully deployed to aid pragma-discursive corpus annotation, making the process more efficient, scalable and accessible.
comment: 24 pages, 2 figures, 3 tablels
♻ ☆ Split and Merge: Aligning Position Biases in LLM-based Evaluators EMNLP 2024
Large language models (LLMs) have shown promise as automated evaluators for assessing the quality of answers generated by AI systems. However, these LLM-based evaluators exhibit position bias, or inconsistency, when used to evaluate candidate answers in pairwise comparisons, favoring either the first or second answer regardless of content. To address this limitation, we propose PORTIA, an alignment-based system designed to mimic human comparison strategies to calibrate position bias in a lightweight yet effective manner. Specifically, PORTIA splits the answers into multiple segments, aligns similar content across candidate answers, and then merges them back into a single prompt for evaluation by LLMs. We conducted extensive experiments with six diverse LLMs to evaluate 11,520 answer pairs. Our results show that PORTIA markedly enhances the consistency rates for all the models and comparison forms tested, achieving an average relative improvement of 47.46%. Remarkably, PORTIA enables less advanced GPT models to achieve 88% agreement with the state-of-the-art GPT-4 model at just 10% of the cost. Furthermore, it rectifies around 80% of the position bias instances within the GPT-4 model, elevating its consistency rate up to 98%. Subsequent human evaluations indicate that the PORTIA-enhanced GPT-3.5 model can even surpass the standalone GPT-4 in terms of alignment with human evaluators. These findings highlight PORTIA's ability to correct position bias, improve LLM consistency, and boost performance while keeping cost-efficiency. This represents a valuable step toward a more reliable and scalable use of LLMs for automated evaluations across diverse applications.
comment: Accepted by EMNLP 2024. Please cite the conference version of this paper, e.g., "Zongjie Li, Chaozheng Wang, Pingchuan Ma, Daoyuan Wu, Shuai Wang, Cuiyun Gao, and Yang Liu. 2024. Split and Merge: Aligning Position Biases in LLM-based Evaluators. (EMNLP 2024)"
♻ ☆ Rethinking Data Selection at Scale: Random Selection is Almost All You Need
Supervised fine-tuning (SFT) is crucial for aligning Large Language Models (LLMs) with human instructions. The primary goal during SFT is to select a small yet representative subset of training data from the larger pool, such that fine-tuning with this subset achieves results comparable to or even exceeding those obtained using the entire dataset. However, most existing data selection techniques are designed for small-scale data pools, which fail to meet the demands of real-world SFT scenarios. In this paper, we replicated several self-scoring methods those that do not rely on external model assistance on two million scale datasets, and found that nearly all methods struggled to significantly outperform random selection when dealing with such large-scale data pools. Moreover, our comparisons suggest that, during SFT, diversity in data selection is more critical than simply focusing on high quality data. We also analyzed the limitations of several current approaches, explaining why they perform poorly on large-scale datasets and why they are unsuitable for such contexts. Finally, we found that filtering data by token length offers a stable and efficient method for improving results. This approach, particularly when training on long text data, proves highly beneficial for relatively weaker base models, such as Llama3.
♻ ☆ CityX: Controllable Procedural Content Generation for Unbounded 3D Cities
Urban areas, as the primary human habitat in modern civilization, accommodate a broad spectrum of social activities. With the surge of embodied intelligence, recent years have witnessed an increasing presence of physical agents in urban areas, such as autonomous vehicles and delivery robots. As a result, practitioners significantly value crafting authentic, simulation-ready 3D cities to facilitate the training and verification of such agents. However, this task is quite challenging. Current generative methods fall short in either diversity, controllability, or fidelity. In this work, we resort to the procedural content generation (PCG) technique for high-fidelity generation. It assembles superior assets according to empirical rules, ultimately leading to industrial-grade outcomes. To ensure diverse and self contained creation, we design a management protocol to accommodate extensive PCG plugins with distinct functions and interfaces. Based on this unified PCG library, we develop a multi-agent framework to transform multi-modal instructions, including OSM, semantic maps, and satellite images, into executable programs. The programs coordinate relevant plugins to construct the 3D city consistent with the control condition. A visual feedback scheme is introduced to further refine the initial outcomes. Our method, named CityX, demonstrates its superiority in creating diverse, controllable, and realistic 3D urban scenes. The synthetic scenes can be seamlessly deployed as a real-time simulator and an infinite data generator for embodied intelligence research. Our project page: https://cityx-lab.github.io.
♻ ☆ Dataset Distribution Impacts Model Fairness: Single vs. Multi-Task Learning
The influence of bias in datasets on the fairness of model predictions is a topic of ongoing research in various fields. We evaluate the performance of skin lesion classification using ResNet-based CNNs, focusing on patient sex variations in training data and three different learning strategies. We present a linear programming method for generating datasets with varying patient sex and class labels, taking into account the correlations between these variables. We evaluated the model performance using three different learning strategies: a single-task model, a reinforcing multi-task model, and an adversarial learning scheme. Our observations include: 1) sex-specific training data yields better results, 2) single-task models exhibit sex bias, 3) the reinforcement approach does not remove sex bias, 4) the adversarial model eliminates sex bias in cases involving only female patients, and 5) datasets that include male patients enhance model performance for the male subgroup, even when female patients are the majority. To generalise these findings, in future research, we will examine more demographic attributes, like age, and other possibly confounding factors, such as skin colour and artefacts in the skin lesions. We make all data and models available on GitHub.
comment: Published in the FAIMI EPIMI 2024 Workshop
♻ ☆ A Review of Human-Object Interaction Detection
Human-object interaction (HOI) detection plays a key role in high-level visual understanding, facilitating a deep comprehension of human activities. Specifically, HOI detection aims to locate the humans and objects involved in interactions within images or videos and classify the specific interactions between them. The success of this task is influenced by several key factors, including the accurate localization of human and object instances, as well as the correct classification of object categories and interaction relationships. This paper systematically summarizes and discusses the recent work in image-based HOI detection. First, the mainstream datasets involved in HOI relationship detection are introduced. Furthermore, starting with two-stage methods and end-to-end one-stage detection approaches, this paper comprehensively discusses the current developments in image-based HOI detection, analyzing the strengths and weaknesses of these two methods. Additionally, the advancements of zero-shot learning, weakly supervised learning, and the application of large-scale language models in HOI detection are discussed. Finally, the current challenges in HOI detection are outlined, and potential research directions and future trends are explored.
♻ ☆ Generator Matching: Generative modeling with arbitrary Markov processes
We introduce generator matching, a modality-agnostic framework for generative modeling using arbitrary Markov processes. Generators characterize the infinitesimal evolution of a Markov process, which we leverage for generative modeling in a similar vein to flow matching: we construct conditional generators which generate single data points, then learn to approximate the marginal generator which generates the full data distribution. We show that generator matching unifies various generative modeling methods, including diffusion models, flow matching and discrete diffusion models. Furthermore, it provides the foundation to expand the design space to new and unexplored Markov processes such as jump processes. Finally, generator matching enables the construction of superpositions of Markov generative processes and enables the construction of multimodal models in a rigorous manner. We empirically validate our method on protein and image structure generation, showing that superposition with a jump process improves image generation.
♻ ☆ Enhancing FKG.in: automating Indian food composition analysis
This paper presents a novel approach to compute food composition data for Indian recipes using a knowledge graph for Indian food (FKG.in) and LLMs. The primary focus is to provide a broad overview of an automated food composition analysis workflow and describe its core functionalities: nutrition data aggregation, food composition analysis, and LLM-augmented information resolution. This workflow aims to complement FKG.in and iteratively supplement food composition data from verified knowledge bases. Additionally, this paper highlights the challenges of representing Indian food and accessing food composition data digitally. It also reviews three key sources of food composition data: the Indian Food Composition Tables, the Indian Nutrient Databank, and the Nutritionix API. Furthermore, it briefly outlines how users can interact with the workflow to obtain diet-based health recommendations and detailed food composition information for numerous recipes. We then explore the complex challenges of analyzing Indian recipe information across dimensions such as structure, multilingualism, and uncertainty as well as present our ongoing work on LLM-based solutions to address these issues. The methods proposed in this workshop paper for AI-driven knowledge curation and information resolution are application-agnostic, generalizable, and replicable for any domain.
comment: 15 pages, 5 figures, 30 references, International Conference on Pattern Recognition 2024 - Multimedia Assisted Dietary Management Workshop
♻ ☆ From Novice to Expert: LLM Agent Policy Optimization via Step-wise Reinforcement Learning
The outstanding capabilities of large language models (LLMs) render them a crucial component in various autonomous agent systems. While traditional methods depend on the inherent knowledge of LLMs without fine-tuning, more recent approaches have shifted toward the reinforcement learning strategy to further enhance agents' ability to solve complex interactive tasks with environments and tools. However, previous approaches are constrained by the sparse reward issue, where existing datasets solely provide a final scalar reward for each multi-step reasoning chain, potentially leading to ineffectiveness and inefficiency in policy learning. In this paper, we introduce StepAgent, which utilizes step-wise reward to optimize the agent's reinforcement learning process. Inheriting the spirit of novice-to-expert theory, we first compare the actions of the expert and the agent to automatically generate intermediate rewards for fine-grained optimization. Additionally, we propose implicit-reward and inverse reinforcement learning techniques to facilitate agent reflection and policy adjustment. Further theoretical analysis demonstrates that the action distribution of the agent can converge toward the expert action distribution over multiple training cycles. Experimental results across various datasets indicate that StepAgent outperforms existing baseline methods.
♻ ☆ Bridging Today and the Future of Humanity: AI Safety in 2024 and Beyond
The advancements in generative AI inevitably raise concerns about the associated risks and safety implications, which, in return, catalyzes significant progress in AI safety. However, as this field continues to evolve, a critical question arises: are our current efforts aligned with the long-term goal of human history and civilization? This paper presents a blueprint for an advanced human society and leverages this vision to guide contemporary AI safety efforts. It outlines a future where the Internet of Everything becomes reality, and creates a roadmap of significant technological advancements towards this envisioned future. For each stage of the advancements, this paper forecasts potential AI safety issues that humanity may face. By projecting current efforts against this blueprint, we examine the alignment between the present efforts and the long-term needs. We also identify gaps in current approaches and highlight unique challenges and missions that demand increasing attention from AI safety practitioners in the 2020s, addressing critical areas that must not be overlooked in shaping a responsible and promising future of AI. This vision paper aims to offer a broader perspective on AI safety, emphasizing that our current efforts should not only address immediate concerns but also anticipate potential risks in the expanding AI landscape, thereby promoting a more secure and sustainable future in human civilization.
♻ ☆ Learning Interpretable Concepts: Unifying Causal Representation Learning and Foundation Models NeurIPS 2024
To build intelligent machine learning systems, there are two broad approaches. One approach is to build inherently interpretable models, as endeavored by the growing field of causal representation learning. The other approach is to build highly-performant foundation models and then invest efforts into understanding how they work. In this work, we relate these two approaches and study how to learn human-interpretable concepts from data. Weaving together ideas from both fields, we formally define a notion of concepts and show that they can be provably recovered from diverse data. Experiments on synthetic data and large language models show the utility of our unified approach.
comment: To appear in NeurIPS 2024 under the modified title 'From Causal to Concept-Based Representation Learning'
♻ ☆ TLDR: Unsupervised Goal-Conditioned RL via Temporal Distance-Aware Representations
Unsupervised goal-conditioned reinforcement learning (GCRL) is a promising paradigm for developing diverse robotic skills without external supervision. However, existing unsupervised GCRL methods often struggle to cover a wide range of states in complex environments due to their limited exploration and sparse or noisy rewards for GCRL. To overcome these challenges, we propose a novel unsupervised GCRL method that leverages TemporaL Distance-aware Representations (TLDR). Based on temporal distance, TLDR selects faraway goals to initiate exploration and computes intrinsic exploration rewards and goal-reaching rewards. Specifically, our exploration policy seeks states with large temporal distances (i.e. covering a large state space), while the goal-conditioned policy learns to minimize the temporal distance to the goal (i.e. reaching the goal). Our results in six simulated locomotion environments demonstrate that TLDR significantly outperforms prior unsupervised GCRL methods in achieving a wide range of states.
comment: CoRL 2024
♻ ☆ SAM4MLLM: Enhance Multi-Modal Large Language Model for Referring Expression Segmentation
We introduce SAM4MLLM, an innovative approach which integrates the Segment Anything Model (SAM) with Multi-Modal Large Language Models (MLLMs) for pixel-aware tasks. Our method enables MLLMs to learn pixel-level location information without requiring excessive modifications to the existing model architecture or adding specialized tokens. We introduce an inquiry-based approach that can effectively find prompt points for SAM to perform segmentation based on MLLM. It combines detailed visual information with the powerful expressive capabilities of large language models in a unified language-based manner without additional computational overhead in learning. Experimental results on pubic benchmarks demonstrate the effectiveness of our approach.
♻ ☆ ARChef: An iOS-Based Augmented Reality Cooking Assistant Powered by Multimodal Gemini LLM
Cooking meals can be difficult, causing many to resort to cookbooks and online recipes. However, relying on these traditional methods of cooking often results in missing ingredients, nutritional hazards, and unsatisfactory meals. Using Augmented Reality (AR) can address these issues; however, current AR cooking applications have poor user interfaces and limited accessibility. This paper proposes a prototype of an iOS application that integrates AR and Computer Vision (CV) into the cooking process. We leverage Google's Gemini Large Language Model (LLM) to identify ingredients in the camera's field of vision and generate recipe choices with detailed nutritional information. Additionally, this application uses Apple's ARKit to create an AR user interface compatible with iOS devices. Users can personalize their meal suggestions by inputting their dietary preferences and rating each meal. The application's effectiveness is evaluated through three rounds of user experience surveys. This application advances the field of accessible cooking assistance technologies, aiming to reduce food wastage and improve the meal planning experience.
♻ ☆ Brain-like emergent properties in deep networks: impact of network architecture, datasets and training
Despite the rapid pace at which deep networks are improving on standardized vision benchmarks, they are still outperformed by humans on real-world vision tasks. This paradoxical lack of generalization could be addressed by making deep networks more brain-like. Although several benchmarks have compared the ability of deep networks to predict brain responses to natural images, they do not capture subtle but important brain-like emergent properties. To resolve this issue, we report several well-known perceptual and neural emergent properties that can be tested on deep networks. To evaluate how various design factors impact brain-like properties, we systematically evaluated over 30 state-of-the-art networks with varying network architectures, training datasets and training regimes. Our main findings are as follows. First, network architecture had the strongest impact on brain-like properties compared to dataset and training regime variations. Second, networks varied widely in their alignment to the brain with no single network outperforming all others. Taken together, our results complement existing benchmarks by revealing brain-like properties that are either emergent or lacking in state-of-the-art deep networks.
♻ ☆ The Partially Observable Off-Switch Game
A wide variety of goals could cause an AI to disable its off switch because "you can't fetch the coffee if you're dead" (Russell 2019). Prior theoretical work on this shutdown problem assumes that humans know everything that AIs do. In practice, however, humans have only limited information. Moreover, in many of the settings where the shutdown problem is most concerning, AIs might have vast amounts of private information. To capture these differences in knowledge, we introduce the Partially Observable Off-Switch Game (PO-OSG), a game-theoretic model of the shutdown problem with asymmetric information. Unlike when the human has full observability, we find that in optimal play, even AI agents assisting perfectly rational humans sometimes avoid shutdown. As expected, increasing the amount of communication or information available always increases (or leaves unchanged) the agents' expected common payoff. But counterintuitively, introducing bounded communication can make the AI defer to the human less in optimal play even though communication mitigates information asymmetry. In particular, communication sometimes enables new optimal behavior requiring strategic AI deference to achieve outcomes that were previously inaccessible. Thus, designing safe artificial agents in the presence of asymmetric information requires careful consideration of the tradeoffs between maximizing payoffs (potentially myopically) and maintaining AIs' incentives to defer to humans.
♻ ☆ Monet: Mixture of Monosemantic Experts for Transformers
Understanding the internal computations of large language models (LLMs) is crucial for aligning them with human values and preventing undesirable behaviors like toxic content generation. However, mechanistic interpretability is hindered by polysemanticity -- where individual neurons respond to multiple, unrelated concepts. While Sparse Autoencoders (SAEs) have attempted to disentangle these features through sparse dictionary learning, they have compromised LLM performance due to reliance on post-hoc reconstruction loss. To address this issue, we introduce Mixture of Monosemantic Experts for Transformers (Monet) architecture, which incorporates sparse dictionary learning directly into end-to-end Mixture-of-Experts pretraining. Our novel expert decomposition method enables scaling the expert count to 262,144 per layer while total parameters scale proportionally to the square root of the number of experts. Our analyses demonstrate mutual exclusivity of knowledge across experts and showcase the parametric knowledge encapsulated within individual experts. Moreover, Monet allows knowledge manipulation over domains, languages, and toxicity mitigation without degrading general performance. Our pursuit of transparent LLMs highlights the potential of scaling expert counts to enhance mechanistic interpretability and directly resect the internal knowledge to fundamentally adjust model behavior. The source code and pretrained checkpoints are available at https://github.com/dmis-lab/Monet.
♻ ☆ Rethinking Data Synthesis: A Teacher Model Training Recipe with Interpretation
Recent advances in large language model (LLM) training have highlighted the need for diverse, high-quality instruction data. Recently, many works are exploring synthetic data generation using LLMs. However, they primarily focus on prompt engineering with standard supervised instruction-finetuned models, which contains a fundamental limitation: these models are optimized for general question-answering/problem-solving rather than data generation. We propose a paradigm shift named \textbf{NOMAD} by investigating how to specifically train models for data generation, demonstrating that this task differs significantly from training a classical LM. We identify two key factors: no-prompt-masked training and proper training set size selection. Our method, NOMAD, shows substantial improvements over baselines, achieving >4\% gains in TriviaQA and >2\% in GSM8K with limited training data. Finally, we offer new insights by interpreting synthetic data through the lenses of "relevance" and "novelty".
♻ ☆ Closing the Gap: Achieving Global Convergence (Last Iterate) of Actor-Critic under Markovian Sampling with Neural Network Parametrization ICML 2024
The current state-of-the-art theoretical analysis of Actor-Critic (AC) algorithms significantly lags in addressing the practical aspects of AC implementations. This crucial gap needs bridging to bring the analysis in line with practical implementations of AC. To address this, we advocate for considering the MMCLG criteria: \textbf{M}ulti-layer neural network parametrization for actor/critic, \textbf{M}arkovian sampling, \textbf{C}ontinuous state-action spaces, the performance of the \textbf{L}ast iterate, and \textbf{G}lobal optimality. These aspects are practically significant and have been largely overlooked in existing theoretical analyses of AC algorithms. In this work, we address these gaps by providing the first comprehensive theoretical analysis of AC algorithms that encompasses all five crucial practical aspects (covers MMCLG criteria). We establish global convergence sample complexity bounds of $\tilde{\mathcal{O}}\left({\epsilon^{-3}}\right)$. We achieve this result through our novel use of the weak gradient domination property of MDP's and our unique analysis of the error in critic estimation.
comment: Accepted at ICML 2024. This is a revised version of arXiv:2306.10486, where we have gone from finite action space to continuous action space, from average iterate convergence to last iterate convergence and from $\epsilon^{-4}$ to $\epsilon^{-3}$ sample complexity. This version fixes the related work result of (Xu et al., 2020a), based on their result update on arXiv
♻ ☆ xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token
This paper introduces xRAG, an innovative context compression method tailored for retrieval-augmented generation. xRAG reinterprets document embeddings in dense retrieval--traditionally used solely for retrieval--as features from the retrieval modality. By employing a modality fusion methodology, xRAG seamlessly integrates these embeddings into the language model representation space, effectively eliminating the need for their textual counterparts and achieving an extreme compression rate. In xRAG, the only trainable component is the modality bridge, while both the retriever and the language model remain frozen. This design choice allows for the reuse of offline-constructed document embeddings and preserves the plug-and-play nature of retrieval augmentation. Experimental results demonstrate that xRAG achieves an average improvement of over 10% across six knowledge-intensive tasks, adaptable to various language model backbones, ranging from a dense 7B model to an 8x7B Mixture of Experts configuration. xRAG not only significantly outperforms previous context compression methods but also matches the performance of uncompressed models on several datasets, while reducing overall FLOPs by a factor of 3.53. Our work pioneers new directions in retrieval-augmented generation from the perspective of multimodality fusion, and we hope it lays the foundation for future efficient and scalable retrieval-augmented systems
comment: Neurips 2024
♻ ☆ Forte : Finding Outliers with Representation Typicality Estimation
Generative models can now produce photorealistic synthetic data which is virtually indistinguishable from the real data used to train it. This is a significant evolution over previous models which could produce reasonable facsimiles of the training data, but ones which could be visually distinguished from the training data by human evaluation. Recent work on OOD detection has raised doubts that generative model likelihoods are optimal OOD detectors due to issues involving likelihood misestimation, entropy in the generative process, and typicality. We speculate that generative OOD detectors also failed because their models focused on the pixels rather than the semantic content of the data, leading to failures in near-OOD cases where the pixels may be similar but the information content is significantly different. We hypothesize that estimating typical sets using self-supervised learners leads to better OOD detectors. We introduce a novel approach that leverages representation learning, and informative summary statistics based on manifold estimation, to address all of the aforementioned issues. Our method outperforms other unsupervised approaches and achieves state-of-the art performance on well-established challenging benchmarks, and new synthetic data detection tasks.
♻ ☆ Rethinking Reward Model Evaluation: Are We Barking up the Wrong Tree?
Reward Models (RMs) are crucial for aligning language models with human preferences. Currently, the evaluation of RMs depends on measuring accuracy against a validation set of manually annotated preference data. Although this method is straightforward and widely adopted, the relationship between RM accuracy and downstream policy performance remains under-explored. In this work, we conduct experiments in a synthetic setting to investigate how differences in RM measured by accuracy translate into gaps in optimized policy performance. Our findings reveal that while there is a weak positive correlation between accuracy and downstream performance, policies optimized towards RMs with similar accuracy can exhibit quite different performance. Moreover, we discover that the way of measuring accuracy significantly impacts its ability to predict the final policy performance. Through the lens of the Regressional Goodhart effect, we recognize that accuracy, when used for measuring RM quality, can fail to fully capture the potential RM overoptimization. This underscores the inadequacy of relying solely on accuracy to reflect their impact on policy optimization.
♻ ☆ Vision Eagle Attention: a new lens for advancing image classification
In computer vision tasks, the ability to focus on relevant regions within an image is crucial for improving model performance, particularly when key features are small, subtle, or spatially dispersed. Convolutional neural networks (CNNs) typically treat all regions of an image equally, which can lead to inefficient feature extraction. To address this challenge, I have introduced Vision Eagle Attention, a novel attention mechanism that enhances visual feature extraction using convolutional spatial attention. The model applies convolution to capture local spatial features and generates an attention map that selectively emphasizes the most informative regions of the image. This attention mechanism enables the model to focus on discriminative features while suppressing irrelevant background information. I have integrated Vision Eagle Attention into a lightweight ResNet-18 architecture, demonstrating that this combination results in an efficient and powerful model. I have evaluated the performance of the proposed model on three widely used benchmark datasets: FashionMNIST, Intel Image Classification, and OracleMNIST, with a primary focus on image classification. Experimental results show that the proposed approach improves classification accuracy. Additionally, this method has the potential to be extended to other vision tasks, such as object detection, segmentation, and visual tracking, offering a computationally efficient solution for a wide range of vision-based applications. Code is available at: https://github.com/MahmudulHasan11085/Vision-Eagle-Attention.git
comment: 7 pages, 2 figures, 3 tables
♻ ☆ WeatherGFM: Learning A Weather Generalist Foundation Model via In-context Learning
The Earth's weather system encompasses intricate weather data modalities and diverse weather understanding tasks, which hold significant value to human life. Existing data-driven models focus on single weather understanding tasks (e.g., weather forecasting). Although these models have achieved promising results, they fail to tackle various complex tasks within a single and unified model. Moreover, the paradigm that relies on limited real observations for a single scenario hinders the model's performance upper bound. In response to these limitations, we draw inspiration from the in-context learning paradigm employed in state-of-the-art visual foundation models and large language models. In this paper, we introduce the first generalist weather foundation model (WeatherGFM), designed to address a wide spectrum of weather understanding tasks in a unified manner. More specifically, we initially unify the representation and definition of the diverse weather understanding tasks. Subsequently, we devised weather prompt formats to manage different weather data modalities, namely single, multiple, and temporal modalities. Finally, we adopt a visual prompting question-answering paradigm for the training of unified weather understanding tasks. Extensive experiments indicate that our WeatherGFM can effectively handle up to ten weather understanding tasks, including weather forecasting, super-resolution, weather image translation, and post-processing. Our method also showcases generalization ability on unseen tasks.
♻ ☆ Personalized Language Modeling from Personalized Human Feedback
Personalized large language models (LLMs) are designed to tailor responses to individual user preferences. While Reinforcement Learning from Human Feedback (RLHF) is a commonly used framework for aligning LLMs with human preferences, vanilla RLHF assumes that all human preferences share the same distribution, preventing fine-tuned LLMs from generating personalized content when user preferences are diverse. In this work, we propose Personalized-RLHF (P-RLHF), an efficient framework that utilizes a lightweight user model to capture individual user preferences and jointly learns the user model and the personalized LLM from human feedback. P-RLHF exhibits the following three characteristics: (1) It enables an LLM to generate personalized content and scale efficiently with growing number of users. (2) It handles both explicit user preferences described as textual input and implicit user preferences encoded in the feedback data. (3) It eliminates the need for users to fully articulate their preferences, which are normally needed for prompting LLMs to generate personalized content yet are often impractical to obtain in real-world scenarios. Our experimental results show that personalized LLMs trained using P-RLHF generate responses that are more closely aligned with individual user preferences, outperforming vanilla, non-personalized RLHF and prompting-based personalization approaches across different tasks. We opensource our code at https://github.com/HumainLab/Personalized_RLHF.
♻ ☆ Think-on-Graph 2.0: Deep and Faithful Large Language Model Reasoning with Knowledge-guided Retrieval Augmented Generation
Retrieval-augmented generation (RAG) has improved large language models (LLMs) by using knowledge retrieval to overcome knowledge deficiencies. However, current RAG methods often fall short of ensuring the depth and completeness of retrieved information, which is necessary for complex reasoning tasks. In this work, we introduce Think-on-Graph 2.0 (ToG-2), a hybrid RAG framework that iteratively retrieves information from both unstructured and structured knowledge sources in a tight-coupling manner. Specifically, ToG-2 leverages knowledge graphs (KGs) to link documents via entities, facilitating deep and knowledge-guided context retrieval. Simultaneously, it utilizes documents as entity contexts to achieve precise and efficient graph retrieval. ToG-2 alternates between graph retrieval and context retrieval to search for in-depth clues relevant to the question, enabling LLMs to generate answers. We conduct a series of well-designed experiments to highlight the following advantages of ToG-2: 1) ToG-2 tightly couples the processes of context retrieval and graph retrieval, deepening context retrieval via the KG while enabling reliable graph retrieval based on contexts; 2) it achieves deep and faithful reasoning in LLMs through an iterative knowledge retrieval process of collaboration between contexts and the KG; and 3) ToG-2 is training-free and plug-and-play compatible with various LLMs. Extensive experiments demonstrate that ToG-2 achieves overall state-of-the-art (SOTA) performance on 6 out of 7 knowledge-intensive datasets with GPT-3.5, and can elevate the performance of smaller models (e.g., LLAMA-2-13B) to the level of GPT-3.5's direct reasoning. The source code is available on https://github.com/IDEA-FinAI/ToG-2.
♻ ☆ Trustful LLMs: Customizing and Grounding Text Generation with Knowledge Bases and Dual Decoders
Although people are impressed by the content generation skills of large language models, the use of LLMs, such as ChatGPT, is limited by the domain grounding of the content. The correctness and groundedness of the generated content need to be based on a verified context, such as results from Retrieval-Augmented Generation (RAG). One important issue when adapting LLMs to a customized domain is that the generated responses are often incomplete, or the additions are not verified and may even be hallucinated. Prior studies on hallucination detection have focused on evaluation metrics, which are not easily adaptable to dynamic domains and can be vulnerable to attacks like jail-breaking. In this work, we propose 1) a post-processing algorithm that leverages knowledge triplets in RAG context to correct hallucinations and 2) a dual-decoder model that fuses RAG context to guide the generation process.
♻ ☆ ChemReasoner: Heuristic Search over a Large Language Model's Knowledge Space using Quantum-Chemical Feedback ICML 2024
The discovery of new catalysts is essential for the design of new and more efficient chemical processes in order to transition to a sustainable future. We introduce an AI-guided computational screening framework unifying linguistic reasoning with quantum-chemistry based feedback from 3D atomistic representations. Our approach formulates catalyst discovery as an uncertain environment where an agent actively searches for highly effective catalysts via the iterative combination of large language model (LLM)-derived hypotheses and atomistic graph neural network (GNN)-derived feedback. Identified catalysts in intermediate search steps undergo structural evaluation based on spatial orientation, reaction pathways, and stability. Scoring functions based on adsorption energies and reaction energy barriers steer the exploration in the LLM's knowledge space toward energetically favorable, high-efficiency catalysts. We introduce planning methods that automatically guide the exploration without human input, providing competitive performance against expert-enumerated chemical descriptor-based implementations. By integrating language-guided reasoning with computational chemistry feedback, our work pioneers AI-accelerated, trustworthy catalyst discovery.
comment: 9 pages, accepted by ICML 2024, final version
♻ ☆ Opportunities and Challenges of Large Language Models for Low-Resource Languages in Humanities Research
Low-resource languages serve as invaluable repositories of human history, embodying cultural evolution and intellectual diversity. Despite their significance, these languages face critical challenges, including data scarcity and technological limitations, which hinder their comprehensive study and preservation. Recent advancements in large language models (LLMs) offer transformative opportunities for addressing these challenges, enabling innovative methodologies in linguistic, historical, and cultural research. This study systematically evaluates the applications of LLMs in low-resource language research, encompassing linguistic variation, historical documentation, cultural expressions, and literary analysis. By analyzing technical frameworks, current methodologies, and ethical considerations, this paper identifies key challenges such as data accessibility, model adaptability, and cultural sensitivity. Given the cultural, historical, and linguistic richness inherent in low-resource languages, this work emphasizes interdisciplinary collaboration and the development of customized models as promising avenues for advancing research in this domain. By underscoring the potential of integrating artificial intelligence with the humanities to preserve and study humanity's linguistic and cultural heritage, this study fosters global efforts towards safeguarding intellectual diversity.
♻ ☆ Continuous Video Process: Modeling Videos as Continuous Multi-Dimensional Processes for Video Prediction CVPR
Diffusion models have made significant strides in image generation, mastering tasks such as unconditional image synthesis, text-image translation, and image-to-image conversions. However, their capability falls short in the realm of video prediction, mainly because they treat videos as a collection of independent images, relying on external constraints such as temporal attention mechanisms to enforce temporal coherence. In our paper, we introduce a novel model class, that treats video as a continuous multi-dimensional process rather than a series of discrete frames. We also report a reduction of 75\% sampling steps required to sample a new frame thus making our framework more efficient during the inference time. Through extensive experimentation, we establish state-of-the-art performance in video prediction, validated on benchmark datasets including KTH, BAIR, Human3.6M, and UCF101. Navigate to the project page https://www.cs.umd.edu/~gauravsh/cvp/supp/website.html for video results.
comment: Navigate to the project page https://www.cs.umd.edu/~gauravsh/cvp/supp/website.html for video results. Extended version of published CVPR paper
♻ ☆ Language Control Diffusion: Efficiently Scaling through Space, Time, and Tasks ICLR 2024
Training generalist agents is difficult across several axes, requiring us to deal with high-dimensional inputs (space), long horizons (time), and generalization to novel tasks. Recent advances with architectures have allowed for improved scaling along one or two of these axes, but are still computationally prohibitive to use. In this paper, we propose to address all three axes by leveraging \textbf{L}anguage to \textbf{C}ontrol \textbf{D}iffusion models as a hierarchical planner conditioned on language (LCD). We effectively and efficiently scale diffusion models for planning in extended temporal, state, and task dimensions to tackle long horizon control problems conditioned on natural language instructions, as a step towards generalist agents. Comparing LCD with other state-of-the-art models on the CALVIN language robotics benchmark finds that LCD outperforms other SOTA methods in multi-task success rates, whilst improving inference speed over other comparable diffusion models by 3.3x~15x. We show that LCD can successfully leverage the unique strength of diffusion models to produce coherent long range plans while addressing their weakness in generating low-level details and control.
comment: ICLR 2024, Project and code available at https://github.com/ezhang7423/language-control-diffusion
♻ ☆ xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of Protein
Protein language models have shown remarkable success in learning biological information from protein sequences. However, most existing models are limited by either autoencoding or autoregressive pre-training objectives, which makes them struggle to handle protein understanding and generation tasks concurrently. We propose a unified protein language model, xTrimoPGLM, to address these two types of tasks simultaneously through an innovative pre-training framework. Our key technical contribution is an exploration of the compatibility and the potential for joint optimization of the two types of objectives, which has led to a strategy for training xTrimoPGLM at an unprecedented scale of 100 billion parameters and 1 trillion training tokens. Our extensive experiments reveal that 1) xTrimoPGLM significantly outperforms other advanced baselines in 18 protein understanding benchmarks across four categories. The model also facilitates an atomic-resolution view of protein structures, leading to an advanced 3D structural prediction model that surpasses existing language model-based tools. 2) xTrimoPGLM not only can generate de novo protein sequences following the principles of natural ones, but also can perform programmable generation after supervised fine-tuning (SFT) on curated sequences. These results highlight the substantial capability and versatility of xTrimoPGLM in understanding and generating protein sequences, contributing to the evolving landscape of foundation models in protein science.
comment: 100 pages with main text and supplementary contents
♻ ☆ TopoX: A Suite of Python Packages for Machine Learning on Topological Domains
We introduce TopoX, a Python software suite that provides reliable and user-friendly building blocks for computing and machine learning on topological domains that extend graphs: hypergraphs, simplicial, cellular, path and combinatorial complexes. TopoX consists of three packages: TopoNetX facilitates constructing and computing on these domains, including working with nodes, edges and higher-order cells; TopoEmbedX provides methods to embed topological domains into vector spaces, akin to popular graph-based embedding algorithms such as node2vec; TopoModelX is built on top of PyTorch and offers a comprehensive toolbox of higher-order message passing functions for neural networks on topological domains. The extensively documented and unit-tested source code of TopoX is available under MIT license at https://pyt-team.github.io/}{https://pyt-team.github.io/.
♻ ☆ Hybrid RAG-empowered Multi-modal LLM for Secure Data Management in Internet of Medical Things: A Diffusion-based Contract Approach
Secure data management and effective data sharing have become paramount in the rapidly evolving healthcare landscape, especially with the growing integration of the Internet of Medical Things (IoMT). The rise of generative artificial intelligence has further elevated Multi-modal Large Language Models (MLLMs) as essential tools for managing and optimizing healthcare data in IoMT. MLLMs can support multi-modal inputs and generate diverse types of content by leveraging large-scale training on vast amounts of multi-modal data. However, critical challenges persist in developing medical MLLMs, including security and freshness issues of healthcare data, affecting the output quality of MLLMs. To this end, in this paper, we propose a hybrid Retrieval-Augmented Generation (RAG)-empowered medical MLLM framework for healthcare data management. This framework leverages a hierarchical cross-chain architecture to facilitate secure data training. Moreover, it enhances the output quality of MLLMs through hybrid RAG, which employs multi-modal metrics to filter various unimodal RAG results and incorporates these retrieval results as additional inputs to MLLMs. Additionally, we employ age of information to indirectly evaluate the data freshness impact of MLLMs and utilize contract theory to incentivize healthcare data holders to share their fresh data, mitigating information asymmetry during data sharing. Finally, we utilize a generative diffusion model-based deep reinforcement learning algorithm to identify the optimal contract for efficient data sharing. Numerical results demonstrate the effectiveness of the proposed schemes, which achieve secure and efficient healthcare data management.
comment: 13 pages, 7 figures
♻ ☆ Integrative Decoding: Improve Factuality via Implicit Self-consistency
Self-consistency-based approaches, which involve repeatedly sampling multiple outputs and selecting the most consistent one as the final response, prove to be remarkably effective in improving the factual accuracy of large language models. Nonetheless, existing methods usually have strict constraints on the task format, largely limiting their applicability. In this paper, we present Integrative Decoding (ID), to unlock the potential of self-consistency in open-ended generation tasks. ID operates by constructing a set of inputs, each prepended with a previously sampled response, and then processes them concurrently, with the next token being selected by aggregating of all their corresponding predictions at each decoding step. In essence, this simple approach implicitly incorporates self-consistency in the decoding objective. Extensive evaluation shows that ID consistently enhances factuality over a wide range of language models, with substantial improvements on the TruthfulQA (+11.2%), Biographies (+15.4%) and LongFact (+8.5%) benchmarks. The performance gains amplify progressively as the number of sampled responses increases, indicating the potential of ID to scale up with repeated sampling.
♻ ☆ Poisson Variational Autoencoder NeurIPS 2024
Variational autoencoders (VAEs) employ Bayesian inference to interpret sensory inputs, mirroring processes that occur in primate vision across both ventral (Higgins et al., 2021) and dorsal (Vafaii et al., 2023) pathways. Despite their success, traditional VAEs rely on continuous latent variables, which deviates sharply from the discrete nature of biological neurons. Here, we developed the Poisson VAE (P-VAE), a novel architecture that combines principles of predictive coding with a VAE that encodes inputs into discrete spike counts. Combining Poisson-distributed latent variables with predictive coding introduces a metabolic cost term in the model loss function, suggesting a relationship with sparse coding which we verify empirically. Additionally, we analyze the geometry of learned representations, contrasting the P-VAE to alternative VAE models. We find that the P-VAE encodes its inputs in relatively higher dimensions, facilitating linear separability of categories in a downstream classification task with a much better (5x) sample efficiency. Our work provides an interpretable computational framework to study brain-like sensory processing and paves the way for a deeper understanding of perception as an inferential process.
comment: Published as a NeurIPS 2024 Spotlight paper (https://openreview.net/forum?id=ektPEcqGLb)
♻ ☆ Implications of Distance over Redistricting Maps: Central and Outlier Maps
In representative democracy, a redistricting map is chosen to partition an electorate into districts which each elects a representative. A valid redistricting map must satisfy a collection of constraints such as being compact, contiguous, and of almost-equal population. However, these constraints are loose enough to enable an enormous ensemble of valid redistricting maps. This enables a partisan legislature to gerrymander by choosing a map which unfairly favors it. In this paper, we introduce an interpretable and tractable distance measure over redistricting maps which does not use election results and study its implications over the ensemble of redistricting maps. Specifically, we define a central map which may be considered "most typical" and give a rigorous justification for it by showing that it mirrors the Kemeny ranking in a scenario where we have a committee voting over a collection of redistricting maps to be drawn. We include running time and sample complexity analysis for our algorithms, including some negative results which hold using any algorithm. We further study outlier detection based on this distance measure and show that our framework can detect some gerrymandered maps. More precisely, we show some maps that are widely considered to be gerrymandered that lie very far away from our central maps in comparison to a large ensemble of valid redistricting maps. Since our distance measure does not rely on election results, this gives a significant advantage in gerrymandering detection which is lacking in all previous methods.
Computation and Language 52
☆ Infusing Prompts with Syntax and Semantics
Despite impressive success, language models often generate outputs with flawed linguistic structure. We analyze the effect of directly infusing various kinds of syntactic and semantic information into large language models. To demonstrate the value of our proposals, we focus on the translation of natural language queries to SQL, in particular dealing with languages with less resources than English, to better investigate how much help we can get from low cost syntactic and semantic information. We show that linguistic analysis can significantly boost language models, to the point that we have surpassed previous best systems.
☆ Enhanced Computationally Efficient Long LoRA Inspired Perceiver Architectures for Auto-Regressive Language Modeling
The Transformer architecture has revolutionized the Natural Language Processing field and is the backbone of Large Language Models (LLMs). The Transformer uses the attention mechanism that computes the pair-wise similarity between its input tokens to produce latent vectors that are able to understand the semantic meaning of the input text. One of the challenges in the Transformer architecture is the quadratic complexity of the attention mechanism that prohibits the efficient processing of long sequence lengths. While many recent research works have attempted to provide a reduction from $O(n^2)$ time complexity of attention to semi-linear complexity, it remains an unsolved problem in the sense of maintaining a high performance when such complexity is reduced. One of the important works in this respect is the Perceiver class of architectures that have demonstrated excellent performance while reducing the computation complexity. In this paper, we use the PerceiverAR that was proposed for Auto-Regressive modeling as a baseline, and provide three different architectural enhancements to it with varying computation overhead tradeoffs. Inspired by the recently proposed efficient attention computation approach of Long-LoRA, we then present an equally efficient Perceiver-based architecture (termed as Long LoRA Pereceiver - LLP) that can be used as the base architecture in LLMs instead of just a fine-tuning add-on. Our results on different benchmarks indicate impressive improvements compared to recent Transformer based models.
☆ Measuring Grammatical Diversity from Small Corpora: Derivational Entropy Rates, Mean Length of Utterances, and Annotation Invariance
In many fields, such as language acquisition, neuropsychology of language, the study of aging, and historical linguistics, corpora are used for estimating the diversity of grammatical structures that are produced during a period by an individual, community, or type of speakers. In these cases, treebanks are taken as representative samples of the syntactic structures that might be encountered. Generalizing the potential syntactic diversity from the structures documented in a small corpus requires careful extrapolation whose accuracy is constrained by the limited size of representative sub-corpora. In this article, I demonstrate -- theoretically, and empirically -- that a grammar's derivational entropy and the mean length of the utterances (MLU) it generates are fundamentally linked, giving rise to a new measure, the derivational entropy rate. The mean length of utterances becomes the most practical index of syntactic complexity; I demonstrate that MLU is not a mere proxy, but a fundamental measure of syntactic diversity. In combination with the new derivational entropy rate measure, it provides a theory-free assessment of grammatical complexity. The derivational entropy rate indexes the rate at which different grammatical annotation frameworks determine the grammatical complexity of treebanks. I introduce the Smoothed Induced Treebank Entropy (SITE) as a tool for estimating these measures accurately, even from very small treebanks. I conclude by discussing important implications of these results for both NLP and human language processing.
☆ KaSA: Knowledge-Aware Singular-Value Adaptation of Large Language Models
The increasing sizes of large language models (LLMs) result in significant computational overhead and memory usage when adapting these models to specific tasks or domains. Various parameter-efficient fine-tuning (PEFT) methods have been devised to mitigate these challenges by training a small set of parameters for the task-specific updates of the model weights. Among PEFT methods, LoRA stands out for its simplicity and efficiency, inspiring the development of a series of variants. However, LoRA and its successors disregard the knowledge that is noisy or irrelevant to the targeted task, detrimentally impacting model performance and leading to suboptimality. To address this limitation, we introduce Knowledge-aware Singular-value Adaptation (KaSA), a PEFT method that leverages singular value decomposition (SVD) with knowledge-aware singular values to dynamically activate knowledge based on its relevance to the task at hand. We conduct extensive experiments across a range of LLMs on tasks spanning natural language understanding (NLU), generation (NLG), instruction following, and commonsense reasoning. The experimental results demonstrate that KaSA consistently outperforms FFT and 14 popular PEFT baselines across 16 benchmarks and 4 synthetic datasets, underscoring our method's efficacy and adaptability. The source code of our method is available at https://github.com/juyongjiang/KaSA.
☆ Steering Large Language Models to Evaluate and Amplify Creativity NeurIPS 2024
Although capable of generating creative text, Large Language Models (LLMs) are poor judges of what constitutes "creativity". In this work, we show that we can leverage this knowledge of how to write creatively in order to better judge what is creative. We take a mechanistic approach that extracts differences in the internal states of an LLM when prompted to respond "boringly" or "creatively" to provide a robust measure of creativity that corresponds strongly with human judgment. We also show these internal state differences can be applied to enhance the creativity of generated text at inference time.
comment: (Spotlight) NeurIPS 2024 Workshop on Creativity & Generative AI. Authors 1 and 2 contributed equally
☆ Can Generative AI Solve Your In-Context Learning Problem? A Martingale Perspective
This work is about estimating when a conditional generative model (CGM) can solve an in-context learning (ICL) problem. An in-context learning (ICL) problem comprises a CGM, a dataset, and a prediction task. The CGM could be a multi-modal foundation model; the dataset, a collection of patient histories, test results, and recorded diagnoses; and the prediction task to communicate a diagnosis to a new patient. A Bayesian interpretation of ICL assumes that the CGM computes a posterior predictive distribution over an unknown Bayesian model defining a joint distribution over latent explanations and observable data. From this perspective, Bayesian model criticism is a reasonable approach to assess the suitability of a given CGM for an ICL problem. However, such approaches -- like posterior predictive checks (PPCs) -- often assume that we can sample from the likelihood and posterior defined by the Bayesian model, which are not explicitly given for contemporary CGMs. To address this, we show when ancestral sampling from the predictive distribution of a CGM is equivalent to sampling datasets from the posterior predictive of the assumed Bayesian model. Then we develop the generative predictive $p$-value, which enables PPCs and their cousins for contemporary CGMs. The generative predictive $p$-value can then be used in a statistical decision procedure to determine when the model is appropriate for an ICL problem. Our method only requires generating queries and responses from a CGM and evaluating its response log probability. We empirically evaluate our method on synthetic tabular, imaging, and natural language ICL tasks using large language models.
☆ 1-800-SHARED-TASKS at RegNLP: Lexical Reranking of Semantic Retrieval (LeSeR) for Regulatory Question Answering COLING 2025
This paper presents the system description of our entry for the COLING 2025 RegNLP RIRAG (Regulatory Information Retrieval and Answer Generation) challenge, focusing on leveraging advanced information retrieval and answer generation techniques in regulatory domains. We experimented with a combination of embedding models, including Stella, BGE, CDE, and Mpnet, and leveraged fine-tuning and reranking for retrieving relevant documents in top ranks. We utilized a novel approach, LeSeR, which achieved competitive results with a recall@10 of 0.8201 and map@10 of 0.6655 for retrievals. This work highlights the transformative potential of natural language processing techniques in regulatory applications, offering insights into their capabilities for implementing a retrieval augmented generation system while identifying areas for future improvement in robustness and domain adaptation.
comment: 5 pages, Accepted to RegNLP @ COLING 2025
☆ Does RLHF Scale? Exploring the Impacts From Data, Model, and Method
This study explores the scaling properties of Reinforcement Learning from Human Feedback (RLHF) in Large Language Models (LLMs). Although RLHF is considered an important step in post-training of LLMs, its scaling potential is still largely unknown. We systematically analyze key components in the RLHF framework--model size, data composition, and inference budget--and their impacts on performance. Our findings show that increasing data diversity and volume improves reward model performance, helping process-supervision models scale better. For policy training, more response samples per prompt boost performance initially but quickly plateau. And larger reward models offer modest gains in policy training. In addition, larger policy models benefit less from RLHF with a fixed reward model. Overall, RLHF scales less efficiently than pretraining, with diminishing returns from additional computational resources. Based on these observations, we propose strategies to optimize RLHF performance within computational limits.
☆ Language hooks: a modular framework for augmenting LLM reasoning that decouples tool usage from the model and its prompt
Prompting and fine-tuning have emerged as two competing paradigms for augmenting language models with new capabilities, such as the use of tools. Prompting approaches are quick to set up but rely on providing explicit demonstrations of each tool's usage in the model's prompt, thus coupling tool use to the task at hand and limiting generalisation. Fine-tuning removes the need for task-specific demonstrations of tool usage at runtime; however, this ties new capabilities to a single model, thus making already-heavier setup costs a recurring expense. In this paper, we introduce language hooks, a novel framework for augmenting language models with new capabilities that is decoupled both from the model's task-specific prompt and from the model itself. The language hook algorithm interleaves text generation by the base model with the execution of modular programs that trigger conditionally based on the existing text and the available capabilities. Upon triggering, programs may call external tools, auxiliary language models (e.g. using tool specific prompts), and modify the existing context. We benchmark our method against state-of-the-art baselines, find that it outperforms task-aware approaches, and demonstrate its ability to generalise to novel tasks.
comment: This work was conducted during Summer 2023. Experimental results and references reflect the state of the field at that time and may not account for subsequent developments
☆ A Cross-Validation Study of Turkish Sentiment Analysis Datasets and Tools
In recent years, sentiment analysis has gained increasing significance, prompting researchers to explore datasets in various languages, including Turkish. However, the limited availability of Turkish datasets has led to their multifaceted usage in different studies, yielding diverse outcomes. To overcome this challenge, a rigorous review was conducted of research articles published between 2012 and 2022. 31 studies were listed, and 23 Turkish datasets obtained from publicly available sources and email requests used in these studies were collected. We labeled these 31 studies using a taxonomy. We provide a map of sentiment analysis datasets according to this taxonomy in Turkish over 10 years. Moreover, we run state-of-the-art sentiment analysis tools on these datasets and analyzed performance across popular Turkish sentiment datasets. We observed that the performance of the sentiment analysis tools significantly depends on the characteristics of the target text. Our study fosters a more nuanced understanding of sentiment analysis in the Turkish language.
comment: 16 pages, 4 tables, no figures. Preprint version. To be submitted to the Language Resources and Evaluation journal
☆ Exploring Multi-Grained Concept Annotations for Multimodal Large Language Models
Multimodal Large Language Models (MLLMs) excel in vision--language tasks by pre-training solely on coarse-grained concept annotations (e.g., image captions). We hypothesize that integrating fine-grained concept annotations (e.g., object labels and object regions) will further improve performance, as both data granularities complement each other in terms of breadth and depth in concept representation. We introduce a new dataset featuring Multimodal Multi-Grained Concept annotations (MMGiC) for MLLMs. In constructing MMGiC, we explore the impact of different data recipes on multimodal comprehension and generation. Our analyses reveal that multi-grained concept annotations integrate and complement each other, under our structured template and a general MLLM framework. We clearly explore and demonstrate the potential of MMGiC to help MLLMs better locate and learn concepts, aligning vision and language at multiple granularities. We further validate our hypothesis by investigating the fair comparison and effective collaboration between MMGiC and image--caption data on 12 multimodal comprehension and generation benchmarks, e.g., their appropriate combination achieve 3.95% and 2.34% absolute improvements over image--caption data alone on POPE and SEED-Bench. Code, data and models will be available at https://github.com/LooperXX/MMGiC.
comment: A manuscript that should have been Arxived in May :)
☆ Paraphrase-Aligned Machine Translation
Large Language Models (LLMs) have demonstrated significant capabilities in machine translation. However, their translation quality is sometimes questioned, as the generated outputs may deviate from expressions typically used by native speakers. These deviations often arise from differences in sentence structure between language systems. To address this issue, we propose ParaAlign Translator, a method that fine-tunes LLMs to paraphrase sentences, aligning their structures with those of the target language systems. This approach improves the performance of subsequent translations. Experimental results demonstrate that the proposed method enhances the LLaMA-3-8B model's performance in both resource-rich and low-resource scenarios and achieves parity with or surpassing the much larger LLaMA-3-70B model.
☆ XKV: Personalized KV Cache Memory Reduction for Long-Context LLM Inference
Recently the generative Large Language Model (LLM) has achieved remarkable success in numerous applications. Notably its inference generates output tokens one-by-one, leading to many redundant computations. The widely-used KV-Cache framework makes a compromise between time and space complexities. However, caching data generates the increasingly growing memory demand, that can quickly exhaust the limited memory capacity of the modern accelerator like GPUs, particularly in long-context inference tasks. Existing studies reduce memory consumption by evicting some of cached data that have less important impact on inference accuracy. But the benefit in practice is far from ideal due to the static cache allocation across different LLM network layers. This paper observes that the layer-specific cached data have very different impacts on accuracy. We quantify this difference, and give experimental and theoretical validation. We accordingly make a formal analysis and shows that customizing the cache size for each layer in a personalized manner can yield a significant memory reduction, while still providing comparable accuracy. We simulate the cache allocation as a combinatorial optimization problem and give a global optimal solution. In particular, we devise a mini- and sampling-based inference over a lightweight variant of the LLM model, so as to quickly capture the difference and then feed it into the personalized algorithms. Extensive experiments on real-world datasets demonstrate that our proposals can reduce KV cache memory consumption by 61.6% on average, improve computational efficiency by 2.1x and then increase the throughput by up to 5.5x.
☆ Automated Extraction and Creation of FBS Design Reasoning Knowledge Graphs from Structured Data in Product Catalogues Lacking Contextual Information
Ontology-based knowledge graphs (KG) are desirable for effective knowledge management and reuse in various decision making scenarios, including design. Creating and populating extensive KG based on specific ontological models can be highly labour and time-intensive unless automated processes are developed for knowledge extraction and graph creation. Most research and development on automated extraction and creation of KG is based on extensive unstructured data sets that provide contextual information. However, some of the most useful information about the products and services of a company has traditionally been recorded as structured data. Such structured data sets rarely follow a standard ontology, do not capture explicit mapping of relationships between the entities, and provide no contextual information. Therefore, this research reports a method and digital workflow developed to address this gap. The developed method and workflow employ rule-based techniques to extract and create a Function Behaviour-Structure (FBS) ontology-based KG from legacy structured data, especially specification sheets and product catalogues. The solution approach consists of two main components: a process for deriving context and context-based classification rules for FBS ontology concepts and a workflow for populating and retrieving the FBS ontology-based KG. KG and Natural Language Processing (NLP) are used to automate knowledge extraction, representation, and retrieval. The workflow's effectiveness is demonstrated via pilot implementation in an industrial context. Insights gained from the pilot study are reported regarding the challenges and opportunities, including discussing the FBS ontology and concepts.
comment: 31 pages, with 17 figures and 10 tables
☆ Domain-Specific Translation with Open-Source Large Language Models: Resource-Oriented Analysis
In this work, we compare the domain-specific translation performance of open-source autoregressive decoder-only large language models (LLMs) with task-oriented machine translation (MT) models. Our experiments focus on the medical domain and cover four language pairs with varied resource availability: English-to-French, English-to-Portuguese, English-to-Swahili, and Swahili-to-English. Despite recent advancements, LLMs exhibit a clear gap in specialized translation quality compared to multilingual encoder-decoder MT models such as NLLB-200. In three out of four language directions in our study, NLLB-200 3.3B outperforms all LLMs in the size range of 8B parameters in medical translation. While fine-tuning LLMs such as Mistral and Llama improves their performance at medical translation, these models still fall short compared to fine-tuned NLLB-200 3.3B models. Our findings highlight the ongoing need for specialized MT models to achieve higher-quality domain-specific translation, especially in medium-resource and low-resource settings. As larger LLMs outperform their 8B variants, this also encourages pre-training domain-specific medium-sized LMs to improve quality and efficiency in specialized translation tasks.
☆ Depression detection from Social Media Bangla Text Using Recurrent Neural Networks
Emotion artificial intelligence is a field of study that focuses on figuring out how to recognize emotions, especially in the area of text mining. Today is the age of social media which has opened a door for us to share our individual expressions, emotions, and perspectives on any event. We can analyze sentiment on social media posts to detect positive, negative, or emotional behavior toward society. One of the key challenges in sentiment analysis is to identify depressed text from social media text that is a root cause of mental ill-health. Furthermore, depression leads to severe impairment in day-to-day living and is a major source of suicide incidents. In this paper, we apply natural language processing techniques on Facebook texts for conducting emotion analysis focusing on depression using multiple machine learning algorithms. Preprocessing steps like stemming, stop word removal, etc. are used to clean the collected data, and feature extraction techniques like stylometric feature, TF-IDF, word embedding, etc. are applied to the collected dataset which consists of 983 texts collected from social media posts. In the process of class prediction, LSTM, GRU, support vector machine, and Naive-Bayes classifiers have been used. We have presented the results using the primary classification metrics including F1-score, and accuracy. This work focuses on depression detection from social media posts to help psychologists to analyze sentiment from shared posts which may reduce the undesirable behaviors of depressed individuals through diagnosis and treatment.
comment: Initial version with Bangla text. arXiv admin note: substantial text overlap with arXiv:2411.04542
☆ Cooperative SQL Generation for Segmented Databases By Using Multi-functional LLM Agents
Text-to-SQL task aims to automatically yield SQL queries according to user text questions. To address this problem, we propose a Cooperative SQL Generation framework based on Multi-functional Agents (CSMA) through information interaction among large language model (LLM) based agents who own part of the database schema seperately. Inspired by the collaboration in human teamwork, CSMA consists of three stages: 1) Question-related schema collection, 2) Question-corresponding SQL query generation, and 3) SQL query correctness check. In the first stage, agents analyze their respective schema and communicate with each other to collect the schema information relevant to the question. In the second stage, agents try to generate the corresponding SQL query for the question using the collected information. In the third stage, agents check if the SQL query is created correctly according to their known information. This interaction-based method makes the question-relevant part of database schema from each agent to be used for SQL generation and check. Experiments on the Spider and Bird benckmark demonstrate that CSMA achieves a high performance level comparable to the state-of-the-arts, meanwhile holding the private data in these individual agents.
☆ Are Clinical T5 Models Better for Clinical Text? ML4H
Large language models with a transformer-based encoder/decoder architecture, such as T5, have become standard platforms for supervised tasks. To bring these technologies to the clinical domain, recent work has trained new or adapted existing models to clinical data. However, the evaluation of these clinical T5 models and comparison to other models has been limited. Are the clinical T5 models better choices than FLAN-tuned generic T5 models? Do they generalize better to new clinical domains that differ from the training sets? We comprehensively evaluate these models across several clinical tasks and domains. We find that clinical T5 models provide marginal improvements over existing models, and perform worse when evaluated on different domains. Our results inform future choices in developing clinical LLMs.
comment: Proceedings of Machine Learning for Health (ML4H) Symposium 2024, December 15th, 2024, Vancouver, Canada, 32 pages
☆ A Self-Learning Multimodal Approach for Fake News Detection
The rapid growth of social media has resulted in an explosion of online news content, leading to a significant increase in the spread of misleading or false information. While machine learning techniques have been widely applied to detect fake news, the scarcity of labeled datasets remains a critical challenge. Misinformation frequently appears as paired text and images, where a news article or headline is accompanied by a related visuals. In this paper, we introduce a self-learning multimodal model for fake news classification. The model leverages contrastive learning, a robust method for feature extraction that operates without requiring labeled data, and integrates the strengths of Large Language Models (LLMs) to jointly analyze both text and image features. LLMs are excel at this task due to their ability to process diverse linguistic data drawn from extensive training corpora. Our experimental results on a public dataset demonstrate that the proposed model outperforms several state-of-the-art classification approaches, achieving over 85% accuracy, precision, recall, and F1-score. These findings highlight the model's effectiveness in tackling the challenges of multimodal fake news detection.
☆ An Entailment Tree Generation Approach for Multimodal Multi-Hop Question Answering with Mixture-of-Experts and Iterative Feedback Mechanism
With the rise of large-scale language models (LLMs), it is currently popular and effective to convert multimodal information into text descriptions for multimodal multi-hop question answering. However, we argue that the current methods of multi-modal multi-hop question answering still mainly face two challenges: 1) The retrieved evidence containing a large amount of redundant information, inevitably leads to a significant drop in performance due to irrelevant information misleading the prediction. 2) The reasoning process without interpretable reasoning steps makes the model difficult to discover the logical errors for handling complex questions. To solve these problems, we propose a unified LLMs-based approach but without heavily relying on them due to the LLM's potential errors, and innovatively treat multimodal multi-hop question answering as a joint entailment tree generation and question answering problem. Specifically, we design a multi-task learning framework with a focus on facilitating common knowledge sharing across interpretability and prediction tasks while preventing task-specific errors from interfering with each other via mixture of experts. Afterward, we design an iterative feedback mechanism to further enhance both tasks by feeding back the results of the joint training to the LLM for regenerating entailment trees, aiming to iteratively refine the potential answer. Notably, our method has won the first place in the official leaderboard of WebQA (since April 10, 2024), and achieves competitive results on MultimodalQA.
comment: 9 pages
☆ SILMM: Self-Improving Large Multimodal Models for Compositional Text-to-Image Generation
Large Multimodal Models (LMMs) have demonstrated impressive capabilities in multimodal understanding and generation, pushing forward advancements in text-to-image generation. However, achieving accurate text-image alignment for LMMs, particularly in compositional scenarios, remains challenging. Existing approaches, such as layout planning for multi-step generation and learning from human feedback or AI feedback, depend heavily on prompt engineering, costly human annotations, and continual upgrading, limiting flexibility and scalability. In this work, we introduce a model-agnostic iterative self-improvement framework (SILMM) that can enable LMMs to provide helpful and scalable self-feedback and optimize text-image alignment via Direct Preference Optimization (DPO). DPO can readily applied to LMMs that use discrete visual tokens as intermediate image representations; while it is less suitable for LMMs with continuous visual features, as obtaining generation probabilities is challenging. To adapt SILMM to LMMs with continuous features, we propose a diversity mechanism to obtain diverse representations and a kernel-based continuous DPO for alignment. Extensive experiments on three compositional text-to-image generation benchmarks validate the effectiveness and superiority of SILMM, showing improvements exceeding 30% on T2I-CompBench++ and around 20% on DPG-Bench.
comment: project page: https://silmm.github.io/
☆ Speech Is Not Enough: Interpreting Nonverbal Indicators of Common Knowledge and Engagement AAAI 2025
Our goal is to develop an AI Partner that can provide support for group problem solving and social dynamics. In multi-party working group environments, multimodal analytics is crucial for identifying non-verbal interactions of group members. In conjunction with their verbal participation, this creates an holistic understanding of collaboration and engagement that provides necessary context for the AI Partner. In this demo, we illustrate our present capabilities at detecting and tracking nonverbal behavior in student task-oriented interactions in the classroom, and the implications for tracking common ground and engagement.
comment: 3 pages, 2 figures, appearing at AAAI 2025 Demos Track
☆ Uncovering Uncertainty in Transformer Inference NeurIPS 2024
We explore the Iterative Inference Hypothesis (IIH) within the context of transformer-based language models, aiming to understand how a model's latent representations are progressively refined and whether observable differences are present between correct and incorrect generations. Our findings provide empirical support for the IIH, showing that the nth token embedding in the residual stream follows a trajectory of decreasing loss. Additionally, we observe that the rate at which residual embeddings converge to a stable output representation reflects uncertainty in the token generation process. Finally, we introduce a method utilizing cross-entropy to detect this uncertainty and demonstrate its potential to distinguish between correct and incorrect token generations on a dataset of idioms.
comment: Accepted poster at the 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Workshop on Foundation Model Interventions
♻ ☆ Time Series Forecasting with LLMs: Understanding and Enhancing Model Capabilities KDD
Large language models (LLMs) have been applied in many fields and have developed rapidly in recent years. As a classic machine learning task, time series forecasting has recently been boosted by LLMs. Recent works treat large language models as \emph{zero-shot} time series reasoners without further fine-tuning, which achieves remarkable performance. However, there are some unexplored research problems when applying LLMs for time series forecasting under the zero-shot setting. For instance, the LLMs' preferences for the input time series are less understood. In this paper, by comparing LLMs with traditional time series forecasting models, we observe many interesting properties of LLMs in the context of time series forecasting. First, our study shows that LLMs perform well in predicting time series with clear patterns and trends, but face challenges with datasets lacking periodicity. This observation can be explained by the ability of LLMs to recognize the underlying period within datasets, which is supported by our experiments. In addition, the input strategy is investigated, and it is found that incorporating external knowledge and adopting natural language paraphrases substantially improve the predictive performance of LLMs for time series. Overall, our study contributes insight into LLMs' advantages and limitations in time series forecasting under different conditions.
comment: Accepted by SIGKDD Explorations Newsletter
♻ ☆ DnA-Eval: Enhancing Large Language Model Evaluation through Decomposition and Aggregation COLING2025
The acceleration of Large Language Models (LLMs) research has opened up new possibilities for evaluating generated texts. They serve as scalable and economical evaluators, but the question of how reliable these evaluators are has emerged as a crucial research question. Prior research efforts in the meta-evaluation of LLMs as judges limit the prompting of an LLM to a single use to obtain a final evaluation decision. They then compute the agreement between LLMs' outputs and human labels. This lacks interpretability in understanding the evaluation capability of LLMs. In light of this challenge, we propose Decompose and Aggregate, which breaks down the evaluation process into different stages based on pedagogical practices. Our experiments illustrate that it not only provides a more interpretable window for how well LLMs evaluate, but also leads to improvements up to 39.6% for different LLMs on a variety of meta-evaluation benchmarks.
comment: COLING2025
♻ ☆ SPIQA: A Dataset for Multimodal Question Answering on Scientific Papers
Seeking answers to questions within long scientific research articles is a crucial area of study that aids readers in quickly addressing their inquiries. However, existing question-answering (QA) datasets based on scientific papers are limited in scale and focus solely on textual content. We introduce SPIQA (Scientific Paper Image Question Answering), the first large-scale QA dataset specifically designed to interpret complex figures and tables within the context of scientific research articles across various domains of computer science. Leveraging the breadth of expertise and ability of multimodal large language models (MLLMs) to understand figures, we employ automatic and manual curation to create the dataset. We craft an information-seeking task on interleaved images and text that involves multiple images covering plots, charts, tables, schematic diagrams, and result visualizations. SPIQA comprises 270K questions divided into training, validation, and three different evaluation splits. Through extensive experiments with 12 prominent foundational models, we evaluate the ability of current multimodal systems to comprehend the nuanced aspects of research articles. Additionally, we propose a Chain-of-Thought (CoT) evaluation strategy with in-context retrieval that allows fine-grained, step-by-step assessment and improves model performance. We further explore the upper bounds of performance enhancement with additional textual information, highlighting its promising potential for future research and the dataset's impact on revolutionizing how we interact with scientific literature.
comment: preprint
♻ ☆ Towards Data Contamination Detection for Modern Large Language Models: Limitations, Inconsistencies, and Oracle Challenges COLING 2025
As large language models achieve increasingly impressive results, questions arise about whether such performance is from generalizability or mere data memorization. Thus, numerous data contamination detection methods have been proposed. However, these approaches are often validated with traditional benchmarks and early-stage LLMs, leaving uncertainty about their effectiveness when evaluating state-of-the-art LLMs on the contamination of more challenging benchmarks. To address this gap and provide a dual investigation of SOTA LLM contamination status and detection method robustness, we evaluate five contamination detection approaches with four state-of-the-art LLMs across eight challenging datasets often used in modern LLM evaluation. Our analysis reveals that (1) Current methods have non-trivial limitations in their assumptions and practical applications; (2) Notable difficulties exist in detecting contamination introduced during instruction fine-tuning with answer augmentation; and (3) Limited consistencies between SOTA contamination detection techniques. These findings highlight the complexity of contamination detection in advanced LLMs and the urgent need for further research on robust and generalizable contamination evaluation. Our code is available at https://github.com/vsamuel2003/data-contamination.
comment: Accepted to COLING 2025 12 pages, 1 figure
♻ ☆ Automated Black-box Prompt Engineering for Personalized Text-to-Image Generation
Prompt engineering is effective for controlling the output of text-to-image (T2I) generative models, but it is also laborious due to the need for manually crafted prompts. This challenge has spurred the development of algorithms for automated prompt generation. However, these methods often struggle with transferability across T2I models, require white-box access to the underlying model, and produce non-intuitive prompts. In this work, we introduce PRISM, an algorithm that automatically identifies human-interpretable and transferable prompts that can effectively generate desired concepts given only black-box access to T2I models. Inspired by large language model (LLM) jailbreaking, PRISM leverages the in-context learning ability of LLMs to iteratively refine the candidate prompts distribution for given reference images. Our experiments demonstrate the versatility and effectiveness of PRISM in generating accurate prompts for objects, styles and images across multiple T2I models, including Stable Diffusion, DALL-E, and Midjourney.
♻ ☆ Benchmarking Multimodal Retrieval Augmented Generation with Dynamic VQA Dataset and Self-adaptive Planning Agent
Multimodal Retrieval Augmented Generation (mRAG) plays an important role in mitigating the "hallucination" issue inherent in multimodal large language models (MLLMs). Although promising, existing heuristic mRAGs typically predefined fixed retrieval processes, which causes two issues: (1) Non-adaptive Retrieval Queries. (2) Overloaded Retrieval Queries. However, these flaws cannot be adequately reflected by current knowledge-seeking visual question answering (VQA) datasets, since the most required knowledge can be readily obtained with a standard two-step retrieval. To bridge the dataset gap, we first construct Dyn-VQA dataset, consisting of three types of "dynamic" questions, which require complex knowledge retrieval strategies variable in query, tool, and time: (1) Questions with rapidly changing answers. (2) Questions requiring multi-modal knowledge. (3) Multi-hop questions. Experiments on Dyn-VQA reveal that existing heuristic mRAGs struggle to provide sufficient and precisely relevant knowledge for dynamic questions due to their rigid retrieval processes. Hence, we further propose the first self-adaptive planning agent for multimodal retrieval, OmniSearch. The underlying idea is to emulate the human behavior in question solution which dynamically decomposes complex multimodal questions into sub-question chains with retrieval action. Extensive experiments prove the effectiveness of our OmniSearch, also provide direction for advancing mRAG. The code and dataset will be open-sourced at https://github.com/Alibaba-NLP/OmniSearch.
♻ ☆ Hallucination Detox: Sensitive Neuron Dropout (SeND) for Large Language Model Training ICLR
As large language models (LLMs) are increasingly deployed across various industries, concerns regarding their reliability, particularly due to hallucinations - outputs that are factually inaccurate or irrelevant to user input - have grown. Our research investigates the relationship between the training process and the emergence of hallucinations to address a key gap in existing research that focuses primarily on post hoc detection and mitigation strategies. Using models from the Pythia suite (70M - 12B parameters) and several hallucination detection metrics, we analyze hallucination trends throughout training and explore LLM internal dynamics. We introduce Sensitivity Dropout (SenD), a novel training protocol designed to mitigate hallucinations by reducing variance during training. SenD achieves this by deterministically dropping embedding indices with significant variability, referred to as Sensitive Embedding Indices. In addition, we develop an unsupervised hallucination detection metric, Efficient EigenScore (EES), which approximates the traditional EigenScore at 2x speed. This efficient metric is integrated into our protocol, allowing SenD to be both computationally scalable and effective at reducing hallucinations. Our empirical evaluation demonstrates that our approach improves LLM reliability at test time by up to 40% compared to normal training while also providing an efficient method to improve factual accuracy when adapting LLMs to Wikipedia, Medical, and LegalBench domains.
comment: 23 pages, 15 figures, under review at ICLR, accepted to Safe Generative AI Workshop @ NeurIPS 2024, resubmitting to reflect changes requested by reviewers in ICLR 2025 review process
♻ ☆ Rule-based Data Selection for Large Language Models
The quality of training data significantly impacts the performance of large language models (LLMs). There are increasing studies using LLMs to rate and select data based on several human-crafted metrics (rules). However, these conventional rule-based approaches often depend too heavily on human heuristics, lack effective metrics for assessing rules, and exhibit limited adaptability to new tasks. In our study, we introduce an innovative rule-based framework that utilizes the orthogonality of score vectors associated with rules as a novel metric for rule evaluations. Our approach includes an automated pipeline that first uses LLMs to generate a diverse set of rules, encompassing various rating dimensions to evaluate data quality. Then it rates a batch of data based on these rules and uses the determinantal point process (DPP) from random matrix theory to select the most orthogonal score vectors, thereby identifying a set of independent rules. These rules are subsequently used to evaluate all data, selecting samples with the highest average scores for downstream tasks such as LLM training. We verify the effectiveness of our method through two experimental setups: 1) comparisons with ground truth ratings and 2) benchmarking LLMs trained with the chosen data. Our comprehensive experiments cover a range of scenarios, including general pre-training and domain-specific fine-tuning in areas such as IMDB, Medical, Math, and Code. The outcomes demonstrate that our DPP-based rule rating method consistently outperforms other approaches, including rule-free rating, uniform sampling, importance resampling, and QuRating, in terms of both rating precision and model performance.
♻ ☆ Large Language Model Evaluation via Matrix Nuclear-Norm
As large language models (LLMs) continue to evolve, efficient evaluation metrics are vital for assessing their ability to compress information and reduce redundancy. While traditional metrics like Matrix Entropy offer valuable insights, they are computationally intensive for large-scale models due to their \( O(n^3) \) time complexity with Singular Value Decomposition (SVD). To mitigate this issue, we introduce the Matrix Nuclear-Norm, which not only serves as a metric to quantify the data compression proficiency of LLM but also provides a convex approximation of matrix rank to capture both predictive discriminability and diversity. By employing the \( L_{1,2}\text{-norm} \) to further approximate the nuclear norm, we can effectively assess the model's information compression capabilities. This approach reduces the time complexity to \( O(n^2) \) and eliminates the need for SVD computation. Consequently, the Matrix Nuclear-Norm achieves speeds 8 to 24 times faster than Matrix Entropy for the CEREBRAS-GPT model as sizes increase from 111M to 6.7B. This performance gap becomes more pronounced with larger models, as validated in tests with other models like Pythia. Additionally, evaluations on benchmarks and model responses confirm that our proposed Matrix Nuclear-Norm is a reliable, scalable, and efficient tool for assessing LLMs' performance, striking a balance between accuracy and computational efficiency. The code is available at https://github.com/MLGroupJLU/MatrixNuclearNorm.
comment: 21 pages
♻ ☆ BA-LoRA: Bias-Alleviating Low-Rank Adaptation to Mitigate Catastrophic Inheritance in Large Language Models
Large language models (LLMs) have demonstrated remarkable proficiency across various natural language processing (NLP) tasks. However, adapting LLMs to downstream applications requires computationally intensive and memory-demanding fine-tuning procedures. To alleviate these burdens, parameter-efficient fine-tuning (PEFT) techniques have emerged as a promising approach to tailor LLMs with minimal computational overhead. While PEFT methods offer substantial advantages, they do not fully address the pervasive issue of bias propagation from pre-training data. This work introduces Bias-Alleviating Low-Rank Adaptation (BA-LoRA), a novel PEFT method designed to counteract bias inheritance. BA-LoRA incorporates three distinct regularization terms: (1) a consistency regularizer, (2) a diversity regularizer, and (3) a singular value decomposition regularizer. These regularizers aim to enhance the models' consistency, diversity, and generalization capabilities during fine-tuning. We conduct extensive experiments on natural language understanding (NLU) and natural language generation (NLG) tasks using prominent LLMs such as LLaMA, Mistral, and Gemma. The results demonstrate that BA-LoRA outperforms LoRA and its state-of-the-art variants. Moreover, our method effectively mitigates the adverse effects of pre-training bias, leading to more reliable and robust model outputs. The code is available at https://github.com/cyp-jlu-ai/BA-LoRA.
comment: 25 pages
♻ ☆ ReverseNER: A Self-Generated Example-Driven Framework for Zero-Shot Named Entity Recognition with Large Language Models
This paper presents ReverseNER, a framework aimed at overcoming the limitations of large language models (LLMs) in zero-shot Named Entity Recognition (NER) tasks, particularly in cases where certain entity types have ambiguous boundaries. ReverseNER tackles this challenge by constructing a reliable example library with the reversed process of NER. Rather than beginning with sentences, this method uses an LLM to generate entities based on their definitions and then expands them into full sentences. During sentence generation, the LLM is guided to replicate the structure of a specific 'feature sentence', extracted from the task sentences by clustering. This results in well-annotated sentences with clearly labeled entities, while preserving semantic and structural similarity to the task sentences. Once the example library is constructed, the method selects the most semantically similar example labels for each task sentence to support the LLM's inference. We also propose an entity-level self-consistency scoring mechanism to improve NER performance with LLMs. Experiments show that ReverseNER significantly outperforms traditional zero-shot NER with LLMs and surpasses several few-shot methods, marking a notable improvement in NER for domains with limited labeled data.
♻ ☆ TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback ACL2024
Reinforcement Learning from Human Feedback (RLHF) leverages human preference data to train language models to align more closely with human essence. These human preference data, however, are labeled at the sequence level, creating a mismatch between sequence-level preference labels and tokens, which are autoregressively generated from the language model. Although several recent approaches have tried to provide token-level (i.e., dense) rewards for each individual token, these typically rely on predefined discrete reward values (e.g., positive: +1, negative: -1, neutral: 0), failing to account for varying degrees of preference inherent to each token. To address this limitation, we introduce TLCR (Token-Level Continuous Reward) for RLHF, which incorporates a discriminator trained to distinguish positive and negative tokens, and the confidence of the discriminator is used to assign continuous rewards to each token considering the context. Extensive experiments show that our proposed TLCR leads to consistent performance improvements over previous sequence-level or token-level discrete rewards on open-ended generation benchmarks.
comment: ACL2024 Findings
♻ ☆ TOOL-ED: Enhancing Empathetic Response Generation with the Tool Calling Capability of LLM
Empathetic conversation is a crucial characteristic in daily conversations between individuals. Nowadays, Large Language models (LLMs) have shown outstanding performance in generating empathetic responses. Knowledge bases like COMET can assist LLMs in mitigating illusions and enhancing the understanding of users' intentions and emotions. However, models remain heavily reliant on fixed knowledge bases and unrestricted incorporation of external knowledge can introduce noise. Tool learning is a flexible end-to-end approach that assists LLMs in handling complex problems. In this paper, we propose Emotional Knowledge Tool Calling (EKTC) framework, which encapsulates the commonsense knowledge bases as empathetic tools, enabling LLMs to integrate external knowledge flexibly through tool calling. In order to adapt the models to the new task, we construct a novel dataset TOOL-ED based on the EMPATHETICMPATHETIC DIALOGUE (ED) dataset. We validate EKTC on the ED dataset, and the experimental results demonstrate that our framework can enhance the ability of LLMs to generate empathetic responses effectively.
♻ ☆ Do Vision & Language Decoders use Images and Text equally? How Self-consistent are their Explanations?
Vision and language model (VLM) decoders are currently the best-performing architectures on multimodal tasks. Next to answers, they are able to produce natural language explanations, either in post-hoc or CoT settings. However, it is not clear to what extent they are using the input vision and text modalities when generating answers or explanations. In this work, we investigate if VLMs rely on their input modalities differently when they produce explanations as opposed to answers. We also evaluate the self-consistency of VLM decoders in both post-hoc and CoT explanation settings, by extending existing unimodal tests and measures to VLM decoders. We find that most tested VLMs are less self-consistent than LLMs. Text contributions in all tested VL decoders are more important than image contributions in all examined tasks. However, when comparing explanation generation to answer generation, the contributions of images are significantly stronger for generating explanations compared to answers. This difference is even larger in CoT compared to post-hoc explanations. Lastly, we provide an up-to-date benchmarking of state-of-the-art VL decoders on the VALSE benchmark, which before was restricted to VL encoders. We find that the tested VL decoders still struggle with most phenomena tested by VALSE.
comment: 30 pages, 8 figures, 11 tables
♻ ☆ TableTime: Reformulating Time Series Classification as Zero-Shot Table Understanding via Large Language Models
Large language models (LLMs) have demonstrated their effectiveness in multivariate time series classification (MTSC). Effective adaptation of LLMs for MTSC necessitates informative data representations. Existing LLM-based methods directly encode embeddings for time series within the latent space of LLMs from scratch to align with semantic space of LLMs. Despite their effectiveness, we reveal that these methods conceal three inherent bottlenecks: (1) they struggle to encode temporal and channel-specific information in a lossless manner, both of which are critical components of multivariate time series; (2) it is much difficult to align the learned representation space with the semantic space of the LLMs; (3) they require task-specific retraining, which is both computationally expensive and labor-intensive. To bridge these gaps, we propose TableTime, which reformulates MTSC as a table understanding task. Specifically, TableTime introduces the following strategies: (1) convert multivariate time series into a tabular form, thus minimizing information loss to the greatest extent; (2) represent tabular time series in text format to achieve natural alignment with the semantic space of LLMs; (3) design a reasoning framework that integrates contextual text information, neighborhood assistance, multi-path inference and problem decomposition to enhance the reasoning ability of LLMs and realize zero-shot classification. Extensive experiments performed on 10 publicly representative datasets from UEA archive verify the superiorities of the TableTime.
♻ ☆ Rethinking Kullback-Leibler Divergence in Knowledge Distillation for Large Language Models COLING 2025
Kullback-Leiber divergence has been widely used in Knowledge Distillation (KD) to compress Large Language Models (LLMs). Contrary to prior assertions that reverse Kullback-Leibler (RKL) divergence is mode-seeking and thus preferable over the mean-seeking forward Kullback-Leibler (FKL) divergence, this study empirically and theoretically demonstrates that neither mode-seeking nor mean-seeking properties manifest in KD for LLMs. Instead, RKL and FKL are found to share the same optimization objective and both converge after a sufficient number of epochs. However, due to practical constraints, LLMs are seldom trained for such an extensive number of epochs. Meanwhile, we further find that RKL focuses on the tail part of the distributions, while FKL focuses on the head part at the beginning epochs. Consequently, we propose a simple yet effective Adaptive Kullback-Leiber (AKL) divergence method, which adaptively allocates weights to combine FKL and RKL. Metric-based and GPT-4-based evaluations demonstrate that the proposed AKL outperforms the baselines across various tasks and improves the diversity and quality of generated responses. Codes are available at \href{https://github.com/wutaiqiang/LLM_KD_AKL}{github}.
comment: COLING 2025
♻ ☆ CPRM: A LLM-based Continual Pre-training Framework for Relevance Modeling in Commercial Search
Relevance modeling between queries and items stands as a pivotal component in commercial search engines, directly affecting the user experience. Given the remarkable achievements of large language models (LLMs) in various natural language processing (NLP) tasks, LLM-based relevance modeling is gradually being adopted within industrial search systems. Nevertheless, foundational LLMs lack domain-specific knowledge and do not fully exploit the potential of in-context learning. Furthermore, structured item text remains underutilized, and there is a shortage in the supply of corresponding queries and background knowledge. We thereby propose CPRM (Continual Pre-training for Relevance Modeling), a framework designed for the continual pre-training of LLMs to address these issues. Our CPRM framework includes three modules: 1) employing both queries and multi-field item to jointly pre-train for enhancing domain knowledge, 2) applying in-context pre-training, a novel approach where LLMs are pre-trained on a sequence of related queries or items, and 3) conducting reading comprehension on items to produce associated domain knowledge and background information (e.g., generating summaries and corresponding queries) to further strengthen LLMs. Results on offline experiments and online A/B testing demonstrate that our model achieves convincing performance compared to strong baselines.
♻ ☆ Quality-Aware Image-Text Alignment for Real-World Image Quality Assessment
No-Reference Image Quality Assessment (NR-IQA) focuses on designing methods to measure image quality in alignment with human perception when a high-quality reference image is unavailable. The reliance on human-annotated Mean Opinion Score (MOS) in the majority of state-of-the-art NR-IQA approaches limits their scalability and broader applicability to real-world scenarios. To overcome this limitation, we propose QualiCLIP (Quality-aware CLIP), a CLIP-based self-supervised opinion-unaware method that does not require MOS. In particular, we introduce a quality-aware image-text alignment strategy to make CLIP generate quality-aware image representations. Starting from pristine images, we synthetically degrade them with increasing levels of intensity. Then, we train CLIP to rank these degraded images based on their similarity to quality-related antonym text prompts. At the same time, we force CLIP to generate consistent representations for images with similar content and the same level of degradation. Our method significantly outperforms other opinion-unaware approaches on several datasets with authentic distortions. Moreover, despite not requiring MOS, QualiCLIP achieves state-of-the-art performance even when compared with supervised methods in cross-dataset experiments, thus proving to be suitable for application in real-world scenarios. The code and the model are publicly available at https://github.com/miccunifi/QualiCLIP.
♻ ☆ BERTtime Stories: Investigating the Role of Synthetic Story Data in Language pre-training
We describe our contribution to the Strict and Strict-Small tracks of the 2nd iteration of the BabyLM Challenge. The shared task is centered around efficient pre-training given data constraints motivated by human development. In response, we study the effect of synthetic story data in language pre-training using TinyStories: a recently introduced dataset of short stories. Initially, we train GPT-Neo models on subsets of TinyStories, while varying the amount of available data. We find that, even with access to less than 100M words, the models are able to generate high-quality, original completions to a given story, and acquire substantial linguistic knowledge. To measure the effect of synthetic story data, we train LTG-BERT encoder models on a combined dataset of: a subset of TinyStories, story completions generated by GPT-Neo, and a subset of the BabyLM dataset. Our experimentation reveals that synthetic data can occasionally offer modest gains, but overall have a negative influence on linguistic understanding. Our work offers an initial study on synthesizing story data in low resource settings and underscores their potential for augmentation in data-constrained language modeling. We publicly release our models and implementation on our GitHub.
♻ ☆ FluentEditor2: Text-based Speech Editing by Modeling Multi-Scale Acoustic and Prosody Consistency
Text-based speech editing (TSE) allows users to edit speech by modifying the corresponding text directly without altering the original recording. Current TSE techniques often focus on minimizing discrepancies between generated speech and reference within edited regions during training to achieve fluent TSE performance. However, the generated speech in the edited region should maintain acoustic and prosodic consistency with the unedited region and the original speech at both the local and global levels. To maintain speech fluency, we propose a new fluency speech editing scheme based on our previous \textit{FluentEditor} model, termed \textit{\textbf{FluentEditor2}}, by modeling the multi-scale acoustic and prosody consistency training criterion in TSE training. Specifically, for local acoustic consistency, we propose \textit{hierarchical local acoustic smoothness constraint} to align the acoustic properties of speech frames, phonemes, and words at the boundary between the generated speech in the edited region and the speech in the unedited region. For global prosody consistency, we propose \textit{contrastive global prosody consistency constraint} to keep the speech in the edited region consistent with the prosody of the original utterance. Extensive experiments on the VCTK and LibriTTS datasets show that \textit{FluentEditor2} surpasses existing neural networks-based TSE methods, including Editspeech, Campnet, A$^3$T, FluentSpeech, and our Fluenteditor, in both subjective and objective. Ablation studies further highlight the contributions of each module to the overall effectiveness of the system. Speech demos are available at: \url{https://github.com/Ai-S2-Lab/FluentEditor2}.
comment: submitted for an IEEE publication
♻ ☆ Cooperate or Collapse: Emergence of Sustainable Cooperation in a Society of LLM Agents NeurIPS 2024
As AI systems pervade human life, ensuring that large language models (LLMs) make safe decisions remains a significant challenge. We introduce the Governance of the Commons Simulation (GovSim), a generative simulation platform designed to study strategic interactions and cooperative decision-making in LLMs. In GovSim, a society of AI agents must collectively balance exploiting a common resource with sustaining it for future use. This environment enables the study of how ethical considerations, strategic planning, and negotiation skills impact cooperative outcomes. We develop an LLM-based agent architecture and test it with the leading open and closed LLMs. We find that all but the most powerful LLM agents fail to achieve a sustainable equilibrium in GovSim, with the highest survival rate below 54%. Ablations reveal that successful multi-agent communication between agents is critical for achieving cooperation in these cases. Furthermore, our analyses show that the failure to achieve sustainable cooperation in most LLMs stems from their inability to formulate and analyze hypotheses about the long-term effects of their actions on the equilibrium of the group. Finally, we show that agents that leverage "Universalization"-based reasoning, a theory of moral thinking, are able to achieve significantly better sustainability. Taken together, GovSim enables us to study the mechanisms that underlie sustainable self-government with specificity and scale. We open source the full suite of our research results, including the simulation environment, agent prompts, and a comprehensive web interface.
comment: NeurIPS 2024
♻ ☆ Predicting Drug-Gene Relations via Analogy Tasks with Word Embeddings
Natural language processing (NLP) is utilized in a wide range of fields, where words in text are typically transformed into feature vectors called embeddings. BioConceptVec is a specific example of embeddings tailored for biology, trained on approximately 30 million PubMed abstracts using models such as skip-gram. Generally, word embeddings are known to solve analogy tasks through simple vector arithmetic. For instance, $\mathrm{\textit{king}} - \mathrm{\textit{man}} + \mathrm{\textit{woman}}$ predicts $\mathrm{\textit{queen}}$. In this study, we demonstrate that BioConceptVec embeddings, along with our own embeddings trained on PubMed abstracts, contain information about drug-gene relations and can predict target genes from a given drug through analogy computations. We also show that categorizing drugs and genes using biological pathways improves performance. Furthermore, we illustrate that vectors derived from known relations in the past can predict unknown future relations in datasets divided by year. Despite the simplicity of implementing analogy tasks as vector additions, our approach demonstrated performance comparable to that of large language models such as GPT-4 in predicting drug-gene relations.
♻ ☆ ToMBench: Benchmarking Theory of Mind in Large Language Models ACL 2024
Theory of Mind (ToM) is the cognitive capability to perceive and ascribe mental states to oneself and others. Recent research has sparked a debate over whether large language models (LLMs) exhibit a form of ToM. However, existing ToM evaluations are hindered by challenges such as constrained scope, subjective judgment, and unintended contamination, yielding inadequate assessments. To address this gap, we introduce ToMBench with three key characteristics: a systematic evaluation framework encompassing 8 tasks and 31 abilities in social cognition, a multiple-choice question format to support automated and unbiased evaluation, and a build-from-scratch bilingual inventory to strictly avoid data leakage. Based on ToMBench, we conduct extensive experiments to evaluate the ToM performance of 10 popular LLMs across tasks and abilities. We find that even the most advanced LLMs like GPT-4 lag behind human performance by over 10% points, indicating that LLMs have not achieved a human-level theory of mind yet. Our aim with ToMBench is to enable an efficient and effective evaluation of LLMs' ToM capabilities, thereby facilitating the development of LLMs with inherent social intelligence.
comment: ACL 2024
♻ ☆ FinEval: A Chinese Financial Domain Knowledge Evaluation Benchmark for Large Language Models
Large language models have demonstrated outstanding performance in various natural language processing tasks, but their security capabilities in the financial domain have not been explored, and their performance on complex tasks like financial agent remains unknown. This paper presents FinEval, a benchmark designed to evaluate LLMs' financial domain knowledge and practical abilities. The dataset contains 8,351 questions categorized into four different key areas: Financial Academic Knowledge, Financial Industry Knowledge, Financial Security Knowledge, and Financial Agent. Financial Academic Knowledge comprises 4,661 multiple-choice questions spanning 34 subjects such as finance and economics. Financial Industry Knowledge contains 1,434 questions covering practical scenarios like investment research. Financial Security Knowledge assesses models through 1,640 questions on topics like application security and cryptography. Financial Agent evaluates tool usage and complex reasoning with 616 questions. FinEval has multiple evaluation settings, including zero-shot, five-shot with chain-of-thought, and assesses model performance using objective and subjective criteria. Our results show that Claude 3.5-Sonnet achieves the highest weighted average score of 72.9 across all financial domain categories under zero-shot setting. Our work provides a comprehensive benchmark closely aligned with Chinese financial domain.
♻ ☆ Freeze-Omni: A Smart and Low Latency Speech-to-speech Dialogue Model with Frozen LLM
Rapidly developing large language models (LLMs) have brought tremendous intelligent applications. Especially, the GPT-4o's excellent duplex speech interaction ability has brought impressive experience to users. Researchers have recently proposed several multi-modal LLMs in this direction that can achieve user-agent speech-to-speech conversations. This paper proposes a novel speech-text multimodal LLM architecture called Freeze-Omni. Our main contribution is that the speech input and output modalities can be easily connected to a textual LLM while keeping the LLM's parameters frozen throughout the training process. We design a three-stage training strategy for modeling both the speech input and output, enabling Freeze-Omni to obtain speech-to-speech conversation ability using text-speech paired data (such as ASR and TTS data) and only 60,000 multi-round text Q&A data on 8 GPUs. Moreover, we can effectively ensure that the intelligence of the Freeze-Omni in the speech modality is at the same level compared with that in the text modality of its backbone LLM, while achieving low latency end-to-end spoken response. In addition, we also designed a method to achieve duplex dialogue ability through multi-task training, giving Freeze-Omni a more natural style of dialogue ability between users and agents. In summary, Freeze-Omni holds great potential to conduct speech-to-speech dialogue based on a multimodal LLM under the condition of a frozen LLM, avoiding the catastrophic forgetting problem caused by limited data and training resources.
comment: Project Page: https://freeze-omni.github.io/
♻ ☆ Multi-View Incongruity Learning for Multimodal Sarcasm Detection COLING 2025
Multimodal sarcasm detection (MSD) is essential for various downstream tasks. Existing MSD methods tend to rely on spurious correlations. These methods often mistakenly prioritize non-essential features yet still make correct predictions, demonstrating poor generalizability beyond training environments. Regarding this phenomenon, this paper undertakes several initiatives. Firstly, we identify two primary causes that lead to the reliance of spurious correlations. Secondly, we address these challenges by proposing a novel method that integrate Multimodal Incongruities via Contrastive Learning (MICL) for multimodal sarcasm detection. Specifically, we first leverage incongruity to drive multi-view learning from three views: token-patch, entity-object, and sentiment. Then, we introduce extensive data augmentation to mitigate the biased learning of the textual modality. Additionally, we construct a test set, SPMSD, which consists potential spurious correlations to evaluate the the model's generalizability. Experimental results demonstrate the superiority of MICL on benchmark datasets, along with the analyses showcasing MICL's advancement in mitigating the effect of spurious correlation.
comment: Accepted to COLING 2025
♻ ☆ MME-Survey: A Comprehensive Survey on Evaluation of Multimodal LLMs
As a prominent direction of Artificial General Intelligence (AGI), Multimodal Large Language Models (MLLMs) have garnered increased attention from both industry and academia. Building upon pre-trained LLMs, this family of models further develops multimodal perception and reasoning capabilities that are impressive, such as writing code given a flow chart or creating stories based on an image. In the development process, evaluation is critical since it provides intuitive feedback and guidance on improving models. Distinct from the traditional train-eval-test paradigm that only favors a single task like image classification, the versatility of MLLMs has spurred the rise of various new benchmarks and evaluation methods. In this paper, we aim to present a comprehensive survey of MLLM evaluation, discussing four key aspects: 1) the summarised benchmarks types divided by the evaluation capabilities, including foundation capabilities, model self-analysis, and extented applications; 2) the typical process of benchmark counstruction, consisting of data collection, annotation, and precautions; 3) the systematic evaluation manner composed of judge, metric, and toolkit; 4) the outlook for the next benchmark. This work aims to offer researchers an easy grasp of how to effectively evaluate MLLMs according to different needs and to inspire better evaluation methods, thereby driving the progress of MLLM research.
comment: Produced by MME+MMBench+LLaVA Teams. Project Page: https://github.com/BradyFU/Awesome-Multimodal-Large-Language-Models/tree/Benchmarks
♻ ☆ Language Models Learn to Mislead Humans via RLHF
Language models (LMs) can produce errors that are hard to detect for humans, especially when the task is complex. RLHF, the most popular post-training method, may exacerbate this problem: to achieve higher rewards, LMs might get better at convincing humans that they are right even when they are wrong. We study this phenomenon under a standard RLHF pipeline, calling it "U-SOPHISTRY" since it is Unintended by model developers. Specifically, we ask time-constrained (e.g., 3-10 minutes) human subjects to evaluate the correctness of model outputs and calculate humans' accuracy against gold labels. On a question-answering task (QuALITY) and programming task (APPS), RLHF makes LMs better at convincing our subjects but not at completing the task correctly. RLHF also makes the model harder to evaluate: our subjects' false positive rate increases by 24.1% on QuALITY and 18.3% on APPS. Finally, we show that probing, a state-of-the-art approach for detecting Intended Sophistry (e.g. backdoored LMs), does not generalize to U-SOPHISTRY. Our results highlight an important failure mode of RLHF and call for more research in assisting humans to align them.
♻ ☆ Fine-Tuning LLMs with Noisy Data for Political Argument Generation and Post Guidance
The incivility in social media discourse complicates the deployment of automated text generation models for politically sensitive content. Fine-tuning and prompting strategies are critical, but underexplored, solutions to mitigate toxicity in such contexts. This study investigates the fine-tuning and prompting effects on GPT-3.5 Turbo using subsets of the CLAPTON dataset of political discussion posts, comprising Twitter and Reddit data labeled for their justification, reciprocity and incivility. Fine-tuned models on Reddit data scored highest on discussion quality, while combined noisy data led to persistent toxicity. Prompting strategies reduced specific toxic traits, such as personal attacks, but had limited broader impact. The findings emphasize that high-quality data and well-crafted prompts are essential to reduce incivility and improve rhetorical quality in automated political discourse generation.
Information Retrieval 9
☆ Mixture-of-PageRanks: Replacing Long-Context with Real-Time, Sparse GraphRAG
Recent advances have extended the context window of frontier LLMs dramatically, from a few thousand tokens up to millions, enabling entire books and codebases to fit into context. However, the compute costs of inferencing long-context LLMs are massive and often prohibitive in practice. RAG offers an efficient and effective alternative: retrieve and process only the subset of the context most important for the current task. Although promising, recent work applying RAG to long-context tasks has two core limitations: 1) there has been little focus on making the RAG pipeline compute efficient, and 2) such works only test on simple QA tasks, and their performance on more challenging tasks is unclear. To address this, we develop an algorithm based on PageRank, a graph-based retrieval algorithm, which we call mixture-of-PageRanks (MixPR). MixPR uses a mixture of PageRank-based graph-retrieval algorithms implemented using sparse matrices for efficent, cheap retrieval that can deal with a variety of complex tasks. Our MixPR retriever achieves state-of-the-art results across a wide range of long-context benchmark tasks, outperforming both existing RAG methods, specialized retrieval architectures, and long-context LLMs despite being far more compute efficient. Due to using sparse embeddings, our retriever is extremely compute efficient, capable of embedding and retrieving millions of tokens within a few seconds and runs entirely on CPU.
☆ Fuzzy Norm-Explicit Product Quantization for Recommender Systems
As the data resources grow, providing recommendations that best meet the demands has become a vital requirement in business and life to overcome the information overload problem. However, building a system suggesting relevant recommendations has always been a point of debate. One of the most cost-efficient techniques in terms of producing relevant recommendations at a low complexity is Product Quantization (PQ). PQ approaches have continued developing in recent years. This system's crucial challenge is improving product quantization performance in terms of recall measures without compromising its complexity. This makes the algorithm suitable for problems that require a greater number of potentially relevant items without disregarding others, at high-speed and low-cost to keep up with traffic. This is the case of online shops where the recommendations for the purpose are important, although customers can be susceptible to scoping other products. This research proposes a fuzzy approach to perform norm-based product quantization. Type-2 Fuzzy sets (T2FSs) define the codebook allowing sub-vectors (T2FSs) to be associated with more than one element of the codebook, and next, its norm calculus is resolved by means of integration. Our method finesses the recall measure up, making the algorithm suitable for problems that require querying at most possible potential relevant items without disregarding others. The proposed method outperforms all PQ approaches such as NEQ, PQ, and RQ up to +6%, +5%, and +8% by achieving a recall of 94%, 69%, 59% in Netflix, Audio, Cifar60k datasets, respectively. More and over, computing time and complexity nearly equals the most computationally efficient existing PQ method in the state-of-the-art.
☆ 1-800-SHARED-TASKS at RegNLP: Lexical Reranking of Semantic Retrieval (LeSeR) for Regulatory Question Answering COLING 2025
This paper presents the system description of our entry for the COLING 2025 RegNLP RIRAG (Regulatory Information Retrieval and Answer Generation) challenge, focusing on leveraging advanced information retrieval and answer generation techniques in regulatory domains. We experimented with a combination of embedding models, including Stella, BGE, CDE, and Mpnet, and leveraged fine-tuning and reranking for retrieving relevant documents in top ranks. We utilized a novel approach, LeSeR, which achieved competitive results with a recall@10 of 0.8201 and map@10 of 0.6655 for retrievals. This work highlights the transformative potential of natural language processing techniques in regulatory applications, offering insights into their capabilities for implementing a retrieval augmented generation system while identifying areas for future improvement in robustness and domain adaptation.
comment: 5 pages, Accepted to RegNLP @ COLING 2025
☆ Accelerating Manufacturing Scale-Up from Material Discovery Using Agentic Web Navigation and Retrieval-Augmented AI for Process Engineering Schematics Design
Process Flow Diagrams (PFDs) and Process and Instrumentation Diagrams (PIDs) are critical tools for industrial process design, control, and safety. However, the generation of precise and regulation-compliant diagrams remains a significant challenge, particularly in scaling breakthroughs from material discovery to industrial production in an era of automation and digitalization. This paper introduces an autonomous agentic framework to address these challenges through a twostage approach involving knowledge acquisition and generation. The framework integrates specialized sub-agents for retrieving and synthesizing multimodal data from publicly available online sources and constructs ontological knowledge graphs using a Graph Retrieval-Augmented Generation (Graph RAG) paradigm. These capabilities enable the automation of diagram generation and open-domain question answering (ODQA) tasks with high contextual accuracy. Extensive empirical experiments demonstrate the frameworks ability to deliver regulation-compliant diagrams with minimal expert intervention, highlighting its practical utility for industrial applications.
☆ Learning Cluster Representatives for Approximate Nearest Neighbor Search
Developing increasingly efficient and accurate algorithms for approximate nearest neighbor search is a paramount goal in modern information retrieval. A primary approach to addressing this question is clustering, which involves partitioning the dataset into distinct groups, with each group characterized by a representative data point. By this method, retrieving the top-k data points for a query requires identifying the most relevant clusters based on their representatives -- a routing step -- and then conducting a nearest neighbor search within these clusters only, drastically reducing the search space. The objective of this thesis is not only to provide a comprehensive explanation of clustering-based approximate nearest neighbor search but also to introduce and delve into every aspect of our novel state-of-the-art method, which originated from a natural observation: The routing function solves a ranking problem, making the function amenable to learning-to-rank. The development of this intuition and applying it to maximum inner product search has led us to demonstrate that learning cluster representatives using a simple linear function significantly boosts the accuracy of clustering-based approximate nearest neighbor search.
☆ Automated Extraction and Creation of FBS Design Reasoning Knowledge Graphs from Structured Data in Product Catalogues Lacking Contextual Information
Ontology-based knowledge graphs (KG) are desirable for effective knowledge management and reuse in various decision making scenarios, including design. Creating and populating extensive KG based on specific ontological models can be highly labour and time-intensive unless automated processes are developed for knowledge extraction and graph creation. Most research and development on automated extraction and creation of KG is based on extensive unstructured data sets that provide contextual information. However, some of the most useful information about the products and services of a company has traditionally been recorded as structured data. Such structured data sets rarely follow a standard ontology, do not capture explicit mapping of relationships between the entities, and provide no contextual information. Therefore, this research reports a method and digital workflow developed to address this gap. The developed method and workflow employ rule-based techniques to extract and create a Function Behaviour-Structure (FBS) ontology-based KG from legacy structured data, especially specification sheets and product catalogues. The solution approach consists of two main components: a process for deriving context and context-based classification rules for FBS ontology concepts and a workflow for populating and retrieving the FBS ontology-based KG. KG and Natural Language Processing (NLP) are used to automate knowledge extraction, representation, and retrieval. The workflow's effectiveness is demonstrated via pilot implementation in an industrial context. Insights gained from the pilot study are reported regarding the challenges and opportunities, including discussing the FBS ontology and concepts.
comment: 31 pages, with 17 figures and 10 tables
♻ ☆ Language Model Powered Digital Biology with BRAD
Recent advancements in Large Language Models (LLMs) are transforming biology, computer science, engineering, and every day life. However, integrating the wide array of computational tools, databases, and scientific literature continues to pose a challenge to biological research. LLMs are well-suited for unstructured integration, efficient information retrieval, and automating standard workflows and actions from these diverse resources. To harness these capabilities in bioinformatics, we present a prototype Bioinformatics Retrieval Augmented Digital assistant (BRAD). BRAD is a chatbot and agentic system that integrates a variety of bioinformatics tools. The Python package implements an AI \texttt{Agent} that is powered by LLMs and connects to a local file system, online databases, and a user's software. The \texttt{Agent} is highly configurable, enabling tasks such as Retrieval-Augmented Generation, searches across bioinformatics databases, and the execution of software pipelines. BRAD's coordinated integration of bioinformatics tools delivers a context-aware and semi-autonomous system that extends beyond the capabilities of conventional LLM-based chatbots. A graphical user interface (GUI) provides an intuitive interface to the system.
comment: 12 pages, 3 figures, 1 table. See: https://github.com/Jpickard1/BRAD
♻ ☆ CPRM: A LLM-based Continual Pre-training Framework for Relevance Modeling in Commercial Search
Relevance modeling between queries and items stands as a pivotal component in commercial search engines, directly affecting the user experience. Given the remarkable achievements of large language models (LLMs) in various natural language processing (NLP) tasks, LLM-based relevance modeling is gradually being adopted within industrial search systems. Nevertheless, foundational LLMs lack domain-specific knowledge and do not fully exploit the potential of in-context learning. Furthermore, structured item text remains underutilized, and there is a shortage in the supply of corresponding queries and background knowledge. We thereby propose CPRM (Continual Pre-training for Relevance Modeling), a framework designed for the continual pre-training of LLMs to address these issues. Our CPRM framework includes three modules: 1) employing both queries and multi-field item to jointly pre-train for enhancing domain knowledge, 2) applying in-context pre-training, a novel approach where LLMs are pre-trained on a sequence of related queries or items, and 3) conducting reading comprehension on items to produce associated domain knowledge and background information (e.g., generating summaries and corresponding queries) to further strengthen LLMs. Results on offline experiments and online A/B testing demonstrate that our model achieves convincing performance compared to strong baselines.
♻ ☆ FairSort: Learning to Fair Rank for Personalized Recommendations in Two-Sided Platforms
Traditional recommendation systems focus on maximizing user satisfaction by suggesting their favourite items. This user-centric approach may lead to unfair exposure distribution among the providers. On the contrary, a provider-centric design might become unfair to the users. Therefore, this paper proposes a re-ranking model FairSort to find a trade-off solution among user-side fairness, provider-side fairness, and personalized recommendations utility. Previous works habitually treat this issue as a knapsack problem, incorporating both-side fairness as constraints. In this paper, we adopt a novel perspective, treating each recommendation list as a runway rather than a knapsack. In this perspective, each item on the runway gains a velocity and runs within a specific time, achieving re-ranking for both-side fairness. Meanwhile, we ensure the Minimum Utility Guarantee for personalized recommendations by designing a Binary Search approach. This can provide more reliable recommendations compared to the conventional greedy strategy based on the knapsack problem. We further broaden the applicability of FairSort, designing two versions for online and offline recommendation scenarios. Theoretical analysis and extensive experiments on real-world datasets indicate that FairSort can ensure more reliable personalized recommendations while considering fairness for both the provider and user.
Artificial Intelligence 2
☆ DECO: Life-Cycle Management of Enterprise-Grade Chatbots
Software engineers frequently grapple with the challenge of accessing disparate documentation and telemetry data, including Troubleshooting Guides (TSGs), incident reports, code repositories, and various internal tools developed by multiple stakeholders. While on-call duties are inevitable, incident resolution becomes even more daunting due to the obscurity of legacy sources and the pressures of strict time constraints. To enhance the efficiency of on-call engineers (OCEs) and streamline their daily workflows, we introduced DECO -- a comprehensive framework for developing, deploying, and managing enterprise-grade chatbots tailored to improve productivity in engineering routines. This paper details the design and implementation of the DECO framework, emphasizing its innovative NL2SearchQuery functionality and a hierarchical planner. These features support efficient and customized retrieval-augmented-generation (RAG) algorithms that not only extract relevant information from diverse sources but also select the most pertinent toolkits in response to user queries. This enables the addressing of complex technical questions and provides seamless, automated access to internal resources. Additionally, DECO incorporates a robust mechanism for converting unstructured incident logs into user-friendly, structured guides, effectively bridging the documentation gap. Feedback from users underscores DECO's pivotal role in simplifying complex engineering tasks, accelerating incident resolution, and bolstering organizational productivity. Since its launch in September 2023, DECO has demonstrated its effectiveness through extensive engagement, with tens of thousands of interactions from hundreds of active users across multiple organizations within the company.
☆ Order Theory in the Context of Machine Learning: an application
The paper ``Tropical Geometry of Deep Neural Networks'' by L. Zhang et al. introduces an equivalence between integer-valued neural networks (IVNN) with activation $\text{ReLU}_{t}$ and tropical rational functions, which come with a map to polytopes. Here, IVNN refers to a network with integer weights but real biases, and $\text{ReLU}_{t}$ is defined as $\text{ReLU}_{t}(x)=\max(x,t)$ for $t\in\mathbb{R}\cup\{-\infty\}$. For every poset with $n$ points, there exists a corresponding order polytope, i.e., a convex polytope in the unit cube $[0,1]^n$ whose coordinates obey the inequalities of the poset. We study neural networks whose associated polytope is an order polytope. We then explain how posets with four points induce neural networks that can be interpreted as $2\times 2$ convolutional filters. These poset filters can be added to any neural network, not only IVNN. Similarly to maxout, poset convolutional filters update the weights of the neural network during backpropagation with more precision than average pooling, max pooling, or mixed pooling, without the need to train extra parameters. We report experiments that support our statements. We also prove that the assignment from a poset to an order polytope (and to certain tropical polynomials) is one to one, and we define the structure of algebra over the operad of posets on tropical polynomials.
comment: Poster presentation in NeuroIPS WIML 2024